diff --git a/.devops/full-cuda.Dockerfile b/.devops/full-cuda.Dockerfile index 2a7da586a..61f671465 100644 --- a/.devops/full-cuda.Dockerfile +++ b/.devops/full-cuda.Dockerfile @@ -6,7 +6,7 @@ ARG CUDA_VERSION=11.7.1 # Target the CUDA build image ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VERSION} -FROM ${BASE_CUDA_DEV_CONTAINER} as build +FROM ${BASE_CUDA_DEV_CONTAINER} AS build # Unless otherwise specified, we make a fat build. ARG CUDA_DOCKER_ARCH=all diff --git a/.devops/full-rocm.Dockerfile b/.devops/full-rocm.Dockerfile index 5cbd2e7a1..680d1cb92 100644 --- a/.devops/full-rocm.Dockerfile +++ b/.devops/full-rocm.Dockerfile @@ -6,7 +6,7 @@ ARG ROCM_VERSION=5.6 # Target the CUDA build image ARG BASE_ROCM_DEV_CONTAINER=rocm/dev-ubuntu-${UBUNTU_VERSION}:${ROCM_VERSION}-complete -FROM ${BASE_ROCM_DEV_CONTAINER} as build +FROM ${BASE_ROCM_DEV_CONTAINER} AS build # Unless otherwise specified, we make a fat build. # List from https://github.com/ggerganov/llama.cpp/pull/1087#issuecomment-1682807878 diff --git a/.devops/full.Dockerfile b/.devops/full.Dockerfile index 6f19afa9c..2a06f82b7 100644 --- a/.devops/full.Dockerfile +++ b/.devops/full.Dockerfile @@ -1,6 +1,6 @@ ARG UBUNTU_VERSION=22.04 -FROM ubuntu:$UBUNTU_VERSION as build +FROM ubuntu:$UBUNTU_VERSION AS build RUN apt-get update && \ apt-get install -y build-essential python3 python3-pip git libcurl4-openssl-dev libgomp1 diff --git a/.devops/llama-cli-cuda.Dockerfile b/.devops/llama-cli-cuda.Dockerfile index bff946cbc..8eda63a89 100644 --- a/.devops/llama-cli-cuda.Dockerfile +++ b/.devops/llama-cli-cuda.Dockerfile @@ -6,7 +6,7 @@ ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VER # Target the CUDA runtime image ARG BASE_CUDA_RUN_CONTAINER=nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION} -FROM ${BASE_CUDA_DEV_CONTAINER} as build +FROM ${BASE_CUDA_DEV_CONTAINER} AS build # Unless otherwise specified, we make a fat build. ARG CUDA_DOCKER_ARCH=all @@ -25,7 +25,7 @@ ENV GGML_CUDA=1 RUN make -j$(nproc) llama-cli -FROM ${BASE_CUDA_RUN_CONTAINER} as runtime +FROM ${BASE_CUDA_RUN_CONTAINER} AS runtime RUN apt-get update && \ apt-get install -y libgomp1 diff --git a/.devops/llama-cli-intel.Dockerfile b/.devops/llama-cli-intel.Dockerfile index bd816f9f5..79dba06a7 100644 --- a/.devops/llama-cli-intel.Dockerfile +++ b/.devops/llama-cli-intel.Dockerfile @@ -1,6 +1,6 @@ ARG ONEAPI_VERSION=2024.1.1-devel-ubuntu22.04 -FROM intel/oneapi-basekit:$ONEAPI_VERSION as build +FROM intel/oneapi-basekit:$ONEAPI_VERSION AS build ARG GGML_SYCL_F16=OFF RUN apt-get update && \ @@ -14,10 +14,12 @@ RUN if [ "${GGML_SYCL_F16}" = "ON" ]; then \ echo "GGML_SYCL_F16 is set" && \ export OPT_SYCL_F16="-DGGML_SYCL_F16=ON"; \ fi && \ - cmake -B build -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx ${OPT_SYCL_F16} && \ + echo "Building with static libs" && \ + cmake -B build -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx \ + ${OPT_SYCL_F16} -DBUILD_SHARED_LIBS=OFF && \ cmake --build build --config Release --target llama-cli -FROM intel/oneapi-basekit:$ONEAPI_VERSION as runtime +FROM intel/oneapi-basekit:$ONEAPI_VERSION AS runtime COPY --from=build /app/build/bin/llama-cli /llama-cli diff --git a/.devops/llama-cli-rocm.Dockerfile b/.devops/llama-cli-rocm.Dockerfile index caa507b08..c3d1ab067 100644 --- a/.devops/llama-cli-rocm.Dockerfile +++ b/.devops/llama-cli-rocm.Dockerfile @@ -6,7 +6,7 @@ ARG ROCM_VERSION=5.6 # Target the CUDA build image ARG BASE_ROCM_DEV_CONTAINER=rocm/dev-ubuntu-${UBUNTU_VERSION}:${ROCM_VERSION}-complete -FROM ${BASE_ROCM_DEV_CONTAINER} as build +FROM ${BASE_ROCM_DEV_CONTAINER} AS build # Unless otherwise specified, we make a fat build. # List from https://github.com/ggerganov/llama.cpp/pull/1087#issuecomment-1682807878 diff --git a/.devops/llama-cli-vulkan.Dockerfile b/.devops/llama-cli-vulkan.Dockerfile index 6155d5881..9b0dad8bf 100644 --- a/.devops/llama-cli-vulkan.Dockerfile +++ b/.devops/llama-cli-vulkan.Dockerfile @@ -1,6 +1,6 @@ ARG UBUNTU_VERSION=jammy -FROM ubuntu:$UBUNTU_VERSION as build +FROM ubuntu:$UBUNTU_VERSION AS build # Install build tools RUN apt update && apt install -y git build-essential cmake wget libgomp1 diff --git a/.devops/llama-cli.Dockerfile b/.devops/llama-cli.Dockerfile index 38382bfc9..7f741aa46 100644 --- a/.devops/llama-cli.Dockerfile +++ b/.devops/llama-cli.Dockerfile @@ -1,6 +1,6 @@ ARG UBUNTU_VERSION=22.04 -FROM ubuntu:$UBUNTU_VERSION as build +FROM ubuntu:$UBUNTU_VERSION AS build RUN apt-get update && \ apt-get install -y build-essential git @@ -11,7 +11,7 @@ COPY . . RUN make -j$(nproc) llama-cli -FROM ubuntu:$UBUNTU_VERSION as runtime +FROM ubuntu:$UBUNTU_VERSION AS runtime RUN apt-get update && \ apt-get install -y libgomp1 diff --git a/.devops/llama-server-cuda.Dockerfile b/.devops/llama-server-cuda.Dockerfile index d7eaa0925..67328cf1c 100644 --- a/.devops/llama-server-cuda.Dockerfile +++ b/.devops/llama-server-cuda.Dockerfile @@ -6,7 +6,7 @@ ARG BASE_CUDA_DEV_CONTAINER=nvidia/cuda:${CUDA_VERSION}-devel-ubuntu${UBUNTU_VER # Target the CUDA runtime image ARG BASE_CUDA_RUN_CONTAINER=nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu${UBUNTU_VERSION} -FROM ${BASE_CUDA_DEV_CONTAINER} as build +FROM ${BASE_CUDA_DEV_CONTAINER} AS build # Unless otherwise specified, we make a fat build. ARG CUDA_DOCKER_ARCH=all @@ -27,7 +27,7 @@ ENV LLAMA_CURL=1 RUN make -j$(nproc) llama-server -FROM ${BASE_CUDA_RUN_CONTAINER} as runtime +FROM ${BASE_CUDA_RUN_CONTAINER} AS runtime RUN apt-get update && \ apt-get install -y libcurl4-openssl-dev libgomp1 curl diff --git a/.devops/llama-server-intel.Dockerfile b/.devops/llama-server-intel.Dockerfile index 8f8fef8c0..f525658dd 100644 --- a/.devops/llama-server-intel.Dockerfile +++ b/.devops/llama-server-intel.Dockerfile @@ -1,6 +1,6 @@ ARG ONEAPI_VERSION=2024.1.1-devel-ubuntu22.04 -FROM intel/oneapi-basekit:$ONEAPI_VERSION as build +FROM intel/oneapi-basekit:$ONEAPI_VERSION AS build ARG GGML_SYCL_F16=OFF RUN apt-get update && \ @@ -14,10 +14,11 @@ RUN if [ "${GGML_SYCL_F16}" = "ON" ]; then \ echo "GGML_SYCL_F16 is set" && \ export OPT_SYCL_F16="-DGGML_SYCL_F16=ON"; \ fi && \ + echo "Building with dynamic libs" && \ cmake -B build -DGGML_SYCL=ON -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DLLAMA_CURL=ON ${OPT_SYCL_F16} && \ cmake --build build --config Release --target llama-server -FROM intel/oneapi-basekit:$ONEAPI_VERSION as runtime +FROM intel/oneapi-basekit:$ONEAPI_VERSION AS runtime RUN apt-get update && \ apt-get install -y libcurl4-openssl-dev curl diff --git a/.devops/llama-server-rocm.Dockerfile b/.devops/llama-server-rocm.Dockerfile index af96c3325..763b4cd3f 100644 --- a/.devops/llama-server-rocm.Dockerfile +++ b/.devops/llama-server-rocm.Dockerfile @@ -6,7 +6,7 @@ ARG ROCM_VERSION=5.6 # Target the CUDA build image ARG BASE_ROCM_DEV_CONTAINER=rocm/dev-ubuntu-${UBUNTU_VERSION}:${ROCM_VERSION}-complete -FROM ${BASE_ROCM_DEV_CONTAINER} as build +FROM ${BASE_ROCM_DEV_CONTAINER} AS build # Unless otherwise specified, we make a fat build. # List from https://github.com/ggerganov/llama.cpp/pull/1087#issuecomment-1682807878 diff --git a/.devops/llama-server-vulkan.Dockerfile b/.devops/llama-server-vulkan.Dockerfile index 49062f84b..13a61ffd8 100644 --- a/.devops/llama-server-vulkan.Dockerfile +++ b/.devops/llama-server-vulkan.Dockerfile @@ -1,6 +1,6 @@ ARG UBUNTU_VERSION=jammy -FROM ubuntu:$UBUNTU_VERSION as build +FROM ubuntu:$UBUNTU_VERSION AS build # Install build tools RUN apt update && apt install -y git build-essential cmake wget diff --git a/.devops/llama-server.Dockerfile b/.devops/llama-server.Dockerfile index a53a5c999..b631d5806 100644 --- a/.devops/llama-server.Dockerfile +++ b/.devops/llama-server.Dockerfile @@ -1,6 +1,6 @@ ARG UBUNTU_VERSION=22.04 -FROM ubuntu:$UBUNTU_VERSION as build +FROM ubuntu:$UBUNTU_VERSION AS build RUN apt-get update && \ apt-get install -y build-essential git libcurl4-openssl-dev curl @@ -13,7 +13,7 @@ ENV LLAMA_CURL=1 RUN make -j$(nproc) llama-server -FROM ubuntu:$UBUNTU_VERSION as runtime +FROM ubuntu:$UBUNTU_VERSION AS runtime RUN apt-get update && \ apt-get install -y libcurl4-openssl-dev libgomp1 diff --git a/.devops/nix/apps.nix b/.devops/nix/apps.nix index 897fce4d3..0ecf19fc5 100644 --- a/.devops/nix/apps.nix +++ b/.devops/nix/apps.nix @@ -10,7 +10,6 @@ "llama-embedding" "llama-server" "llama-quantize" - "llama-train-text-from-scratch" ]; mkApp = name: { type = "app"; diff --git a/.devops/nix/package.nix b/.devops/nix/package.nix index 4ee0d62cb..911c42ecb 100644 --- a/.devops/nix/package.nix +++ b/.devops/nix/package.nix @@ -17,19 +17,19 @@ rocmPackages, vulkan-headers, vulkan-loader, - clblast, + curl, + shaderc, useBlas ? builtins.all (x: !x) [ useCuda useMetalKit - useOpenCL useRocm useVulkan ] && blas.meta.available, useCuda ? config.cudaSupport, - useMetalKit ? stdenv.isAarch64 && stdenv.isDarwin && !useOpenCL, + useMetalKit ? stdenv.isAarch64 && stdenv.isDarwin, useMpi ? false, # Increases the runtime closure size by ~700M - useOpenCL ? false, useRocm ? config.rocmSupport, + enableCurl ? true, useVulkan ? false, llamaVersion ? "0.0.0", # Arbitrary version, substituted by the flake @@ -56,7 +56,6 @@ let ++ lib.optionals useCuda [ "CUDA" ] ++ lib.optionals useMetalKit [ "MetalKit" ] ++ lib.optionals useMpi [ "MPI" ] - ++ lib.optionals useOpenCL [ "OpenCL" ] ++ lib.optionals useRocm [ "ROCm" ] ++ lib.optionals useVulkan [ "Vulkan" ]; @@ -91,6 +90,22 @@ let ps.tiktoken ps.torchWithoutCuda ps.transformers + + # server bench + ps.matplotlib + + # server tests + ps.openai + ps.behave + ps.prometheus-client + + # for examples/pydantic-models-to-grammar-examples.py + ps.docstring-parser + ps.pydantic + + # for scripts/compare-llama-bench.py + ps.gitpython + ps.tabulate ] ); @@ -132,6 +147,7 @@ let vulkanBuildInputs = [ vulkan-headers vulkan-loader + shaderc ]; in @@ -198,19 +214,19 @@ effectiveStdenv.mkDerivation ( optionals effectiveStdenv.isDarwin darwinBuildInputs ++ optionals useCuda cudaBuildInputs ++ optionals useMpi [ mpi ] - ++ optionals useOpenCL [ clblast ] ++ optionals useRocm rocmBuildInputs ++ optionals useBlas [ blas ] - ++ optionals useVulkan vulkanBuildInputs; + ++ optionals useVulkan vulkanBuildInputs + ++ optionals enableCurl [ curl ]; cmakeFlags = [ (cmakeBool "LLAMA_BUILD_SERVER" true) (cmakeBool "BUILD_SHARED_LIBS" (!enableStatic)) (cmakeBool "CMAKE_SKIP_BUILD_RPATH" true) + (cmakeBool "LLAMA_CURL" enableCurl) (cmakeBool "GGML_NATIVE" false) (cmakeBool "GGML_BLAS" useBlas) - (cmakeBool "GGML_CLBLAST" useOpenCL) (cmakeBool "GGML_CUDA" useCuda) (cmakeBool "GGML_HIPBLAS" useRocm) (cmakeBool "GGML_METAL" useMetalKit) @@ -254,7 +270,6 @@ effectiveStdenv.mkDerivation ( useCuda useMetalKit useMpi - useOpenCL useRocm useVulkan ; @@ -281,7 +296,7 @@ effectiveStdenv.mkDerivation ( # Configurations we don't want even the CI to evaluate. Results in the # "unsupported platform" messages. This is mostly a no-op, because # cudaPackages would've refused to evaluate anyway. - badPlatforms = optionals (useCuda || useOpenCL) lib.platforms.darwin; + badPlatforms = optionals useCuda lib.platforms.darwin; # Configurations that are known to result in build failures. Can be # overridden by importing Nixpkgs with `allowBroken = true`. diff --git a/.devops/tools.sh b/.devops/tools.sh index 335382f69..24dcfd350 100755 --- a/.devops/tools.sh +++ b/.devops/tools.sh @@ -8,13 +8,11 @@ arg1="$1" shift if [[ "$arg1" == '--convert' || "$arg1" == '-c' ]]; then - python3 ./convert-hf-to-gguf.py "$@" + python3 ./convert_hf_to_gguf.py "$@" elif [[ "$arg1" == '--quantize' || "$arg1" == '-q' ]]; then ./llama-quantize "$@" elif [[ "$arg1" == '--run' || "$arg1" == '-r' ]]; then ./llama-cli "$@" -elif [[ "$arg1" == '--finetune' || "$arg1" == '-f' ]]; then - ./llama-finetune "$@" elif [[ "$arg1" == '--all-in-one' || "$arg1" == '-a' ]]; then echo "Converting PTH to GGML..." for i in `ls $1/$2/ggml-model-f16.bin*`; do @@ -36,8 +34,6 @@ else echo " ex: --outtype f16 \"/models/7B/\" " echo " --quantize (-q): Optimize with quantization process ggml" echo " ex: \"/models/7B/ggml-model-f16.bin\" \"/models/7B/ggml-model-q4_0.bin\" 2" - echo " --finetune (-f): Run finetune command to create a lora finetune of the model" - echo " See documentation for finetune for command-line parameters" echo " --all-in-one (-a): Execute --convert & --quantize" echo " ex: \"/models/\" 7B" echo " --server (-s): Run a model on the server" diff --git a/.github/ISSUE_TEMPLATE/config.yml b/.github/ISSUE_TEMPLATE/config.yml index c88134dbb..eb8c4b472 100644 --- a/.github/ISSUE_TEMPLATE/config.yml +++ b/.github/ISSUE_TEMPLATE/config.yml @@ -9,5 +9,3 @@ contact_links: - name: Want to contribute? url: https://github.com/ggerganov/llama.cpp/wiki/contribute about: Head to the contribution guide page of the wiki for areas you can help with - - diff --git a/.github/labeler.yml b/.github/labeler.yml index 9c0397d16..89436740d 100644 --- a/.github/labeler.yml +++ b/.github/labeler.yml @@ -16,7 +16,9 @@ SYCL: - any-glob-to-any-file: - ggml/include/ggml-sycl.h - ggml/src/ggml-sycl.cpp - - README-sycl.md + - ggml/src/ggml-sycl/** + - docs/backend/SYCL.md + - examples/sycl/** Nvidia GPU: - changed-files: - any-glob-to-any-file: diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml index 208515287..b9246659a 100644 --- a/.github/workflows/build.yml +++ b/.github/workflows/build.yml @@ -47,7 +47,7 @@ jobs: sysctl -a mkdir build cd build - cmake -DLLAMA_FATAL_WARNINGS=ON -DGGML_METAL_EMBED_LIBRARY=ON -DLLAMA_CURL=ON .. + cmake -DLLAMA_FATAL_WARNINGS=ON -DGGML_METAL_EMBED_LIBRARY=ON -DLLAMA_CURL=ON -DBUILD_SHARED_LIBS=OFF .. cmake --build . --config Release -j $(sysctl -n hw.logicalcpu) - name: Test @@ -105,7 +105,7 @@ jobs: sysctl -a # Metal is disabled due to intermittent failures with Github runners not having a GPU: # https://github.com/ggerganov/llama.cpp/actions/runs/8635935781/job/23674807267#step:5:2313 - cmake -B build -DLLAMA_FATAL_WARNINGS=ON -DGGML_METAL=OFF -DLLAMA_CURL=ON + cmake -B build -DLLAMA_FATAL_WARNINGS=ON -DGGML_METAL=OFF -DLLAMA_CURL=ON -DBUILD_SHARED_LIBS=OFF cmake --build build --config Release -j $(sysctl -n hw.logicalcpu) - name: Test @@ -222,7 +222,7 @@ jobs: run: | mkdir build cd build - cmake .. -DLLAMA_FATAL_WARNINGS=ON -DLLAMA_CURL=ON + cmake .. -DLLAMA_FATAL_WARNINGS=ON -DLLAMA_CURL=ON -DBUILD_SHARED_LIBS=OFF cmake --build . --config Release -j $(nproc) - name: Test @@ -355,8 +355,10 @@ jobs: - name: Dependencies id: depends run: | - sudo apt-get update - sudo apt-get install build-essential libvulkan-dev + wget -qO - https://packages.lunarg.com/lunarg-signing-key-pub.asc | sudo apt-key add - + sudo wget -qO /etc/apt/sources.list.d/lunarg-vulkan-jammy.list https://packages.lunarg.com/vulkan/lunarg-vulkan-jammy.list + sudo apt-get update -y + sudo apt-get install -y build-essential vulkan-sdk - name: Build id: cmake_build @@ -799,6 +801,7 @@ jobs: 7z x "-o${env:RUNNER_TEMP}" $env:RUNNER_TEMP/sde.tar $sde = $(join-path $env:RUNNER_TEMP sde-external-${env:SDE_VERSION}-win/sde.exe) cd build + $env:LLAMA_SKIP_TESTS_SLOW_ON_EMULATOR = 1 & $sde -future -- ctest -L main -C Release --verbose --timeout 900 - name: Determine tag name @@ -857,6 +860,7 @@ jobs: mkdir build cd build cmake .. -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_CUDA=ON -DBUILD_SHARED_LIBS=ON + cmake --build . --config Release -j $((${env:NUMBER_OF_PROCESSORS} - 1)) -t ggml cmake --build . --config Release -j ${env:NUMBER_OF_PROCESSORS} - name: Determine tag name diff --git a/.github/workflows/python-type-check.yml b/.github/workflows/python-type-check.yml new file mode 100644 index 000000000..e5ff5e6d7 --- /dev/null +++ b/.github/workflows/python-type-check.yml @@ -0,0 +1,38 @@ +name: Python Type-Check + +on: + push: + paths: + - '.github/workflows/python-type-check.yml' + - '**.py' + - '**/requirements*.txt' + pull_request: + paths: + - '.github/workflows/python-type-check.yml' + - '**.py' + - '**/requirements*.txt' + +concurrency: + group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }} + cancel-in-progress: true + +jobs: + python-type-check: + runs-on: ubuntu-latest + name: pyright type-check + steps: + - name: Check out source repository + uses: actions/checkout@v4 + - name: Set up Python environment + uses: actions/setup-python@v5 + with: + python-version: "3.11" + - name: Install Python dependencies + # TODO: use a venv + run: pip install -r requirements/requirements-all.txt + - name: Type-check with Pyright + uses: jakebailey/pyright-action@v2 + with: + version: 1.1.370 + level: warning + warnings: true diff --git a/.gitignore b/.gitignore index 177e6a8db..c9b4d9983 100644 --- a/.gitignore +++ b/.gitignore @@ -47,8 +47,10 @@ build* !build-info.cpp.in !build-info.sh !build.zig +!docs/build.md /libllama.so /llama-* +/vulkan-shaders-gen android-ndk-* arm_neon.h cmake-build-* @@ -60,6 +62,11 @@ llama-batched-swift out/ tmp/ +# Deprecated + +/main +/server + # CI !.github/workflows/*.yml @@ -98,13 +105,14 @@ examples/server/*.mjs.hpp # Python -__pycache__ -.venv -/Pipfile -dist -poetry.lock +/.venv +__pycache__/ +*/poetry.lock poetry.toml +# Nix +/result + # Test binaries /tests/test-backend-ops /tests/test-double-float diff --git a/CMakeLists.txt b/CMakeLists.txt index 7a7197282..793709122 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -42,13 +42,14 @@ endif() option(BUILD_SHARED_LIBS "build shared libraries" ${BUILD_SHARED_LIBS_DEFAULT}) +if (WIN32) + add_compile_definitions(_CRT_SECURE_NO_WARNINGS) +endif() + # # option list # -# general -option(LLAMA_CCACHE "llama: use ccache if available" ON) - # debug option(LLAMA_ALL_WARNINGS "llama: enable all compiler warnings" ON) option(LLAMA_ALL_WARNINGS_3RD_PARTY "llama: enable all compiler warnings in 3rd party libs" OFF) @@ -73,20 +74,26 @@ option(LLAMA_CURL "llama: use libcurl to download model from an URL" OFF) include(${CMAKE_CURRENT_SOURCE_DIR}/cmake/build-info.cmake) # override ggml options -set(GGML_CCACHE ${LLAMA_CCACHE}) set(GGML_SANITIZE_THREAD ${LLAMA_SANITIZE_THREAD}) set(GGML_SANITIZE_ADDRESS ${LLAMA_SANITIZE_ADDRESS}) set(GGML_SANITIZE_UNDEFINED ${LLAMA_SANITIZE_UNDEFINED}) set(GGML_ALL_WARNINGS ${LLAMA_ALL_WARNINGS}) set(GGML_FATAL_WARNINGS ${LLAMA_FATAL_WARNINGS}) -set(GGML_LLAMAFILE ON) -set(GGML_CUDA_USE_GRAPHS ON) + +# change the default for these ggml options +if (NOT DEFINED GGML_LLAMAFILE) + set(GGML_LLAMAFILE ON) +endif() + +if (NOT DEFINED GGML_CUDA_USE_GRAPHS) + set(GGML_CUDA_USE_GRAPHS ON) +endif() # transition helpers function (llama_option_depr TYPE OLD NEW) if (${OLD}) message(${TYPE} "${OLD} is deprecated and will be removed in the future.\nUse ${NEW} instead\n") - set(${NEW} ON) + set(${NEW} ON PARENT_SCOPE) endif() endfunction() @@ -96,16 +103,19 @@ llama_option_depr(WARNING LLAMA_KOMPUTE GGML_KOMPUTE) llama_option_depr(WARNING LLAMA_METAL GGML_METAL) llama_option_depr(WARNING LLAMA_METAL_EMBED_LIBRARY GGML_METAL_EMBED_LIBRARY) llama_option_depr(WARNING LLAMA_NATIVE GGML_NATIVE) -llama_option_depr(WARNING LLAMA_OPENMP GGML_OPENMP) llama_option_depr(WARNING LLAMA_RPC GGML_RPC) llama_option_depr(WARNING LLAMA_SYCL GGML_SYCL) llama_option_depr(WARNING LLAMA_SYCL_F16 GGML_SYCL_F16) +llama_option_depr(WARNING LLAMA_CANN GGML_CANN) # # build the library # -add_subdirectory(ggml) +if (NOT TARGET ggml) + add_subdirectory(ggml) + # ... otherwise assume ggml is added by a parent CMakeLists.txt +endif() add_subdirectory(src) # @@ -123,7 +133,16 @@ set(LLAMA_INCLUDE_INSTALL_DIR ${CMAKE_INSTALL_INCLUDEDIR} CACHE PATH "Location o set(LLAMA_LIB_INSTALL_DIR ${CMAKE_INSTALL_LIBDIR} CACHE PATH "Location of library files") set(LLAMA_BIN_INSTALL_DIR ${CMAKE_INSTALL_BINDIR} CACHE PATH "Location of binary files") -get_directory_property(LLAMA_TRANSIENT_DEFINES COMPILE_DEFINITIONS) + +# At the moment some compile definitions are placed within the ggml/src +# directory but not exported on the `ggml` target. This could be improved by +# determining _precisely_ which defines are necessary for the llama-config +# package. +# +get_directory_property(GGML_DIR_DEFINES DIRECTORY ggml/src COMPILE_DEFINITIONS) +get_target_property(GGML_TARGET_DEFINES ggml COMPILE_DEFINITIONS) +set(GGML_TRANSIENT_DEFINES ${GGML_TARGET_DEFINES} ${GGML_DIR_DEFINES}) +get_target_property(GGML_LINK_LIBRARIES ggml LINK_LIBRARIES) set_target_properties(llama PROPERTIES PUBLIC_HEADER ${CMAKE_CURRENT_SOURCE_DIR}/include/llama.h) install(TARGETS llama LIBRARY PUBLIC_HEADER) @@ -146,7 +165,7 @@ install(FILES ${CMAKE_CURRENT_BINARY_DIR}/llama-config.cmake DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/llama) install( - FILES convert-hf-to-gguf.py + FILES convert_hf_to_gguf.py PERMISSIONS OWNER_READ OWNER_WRITE diff --git a/CMakePresets.json b/CMakePresets.json index d69bc0344..bdad38952 100644 --- a/CMakePresets.json +++ b/CMakePresets.json @@ -19,6 +19,7 @@ "cacheVariables": { "CMAKE_EXPORT_COMPILE_COMMANDS": "ON", "CMAKE_CXX_COMPILER": "icx", + "CMAKE_C_COMPILER": "cl", "GGML_SYCL": "ON", "CMAKE_INSTALL_RPATH": "$ORIGIN;$ORIGIN/.." } diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 991d85e49..b688f78ec 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -1,14 +1,28 @@ -# Contributing Guidelines +# Pull requests (for contributors) -## Checklist +- Test your changes: + - Using the commands in the [`tests`](tests) folder. For instance, running the `./tests/test-backend-ops` command tests different backend implementations of the GGML library + - Execute [the full CI locally on your machine](ci/README.md) before publishing +- Please rate the complexity of your PR (i.e. `Review Complexity : Low`, `Review Complexity : Medium`, `Review Complexity : High`). This makes it easier for maintainers to triage the PRs. + - The PR template has a series of review complexity checkboxes `[ ]` that [you can mark as](https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/about-task-lists) `[X]` for your convenience +- If your PR becomes stale, don't hesitate to ping the maintainers in the comments -* Make sure your PR follows the [coding guidelines](https://github.com/ggerganov/llama.cpp/blob/master/README.md#coding-guidelines) -* Test your changes using the commands in the [`tests`](tests) folder. For instance, running the `./tests/test-backend-ops` command tests different backend implementations of the GGML library -* Execute [the full CI locally on your machine](ci/README.md) before publishing +# Pull requests (for collaborators) -## PR formatting +- Squash-merge PRs +- Use the following format for the squashed commit title: ` : (#)`. For example: `utils : fix typo in utils.py (#1234)` +- Optionally, pick a `` from here: https://github.com/ggerganov/llama.cpp/wiki/Modules + +# Coding guidelines + +- Avoid adding third-party dependencies, extra files, extra headers, etc. +- Always consider cross-compatibility with other operating systems and architectures +- Avoid fancy looking modern STL constructs, use basic `for` loops, avoid templates, keep it simple +- There are no strict rules for the code style, but try to follow the patterns in the code (indentation, spaces, etc.). Vertical alignment makes things more readable and easier to batch edit +- Clean-up any trailing whitespaces, use 4 spaces for indentation, brackets on the same line, `void * ptr`, `int & a` +- Naming usually optimizes for common prefix (see https://github.com/ggerganov/ggml/pull/302#discussion_r1243240963) +- Tensors store data in row-major order. We refer to dimension 0 as columns, 1 as rows, 2 as matrices +- Matrix multiplication is unconventional: [`C = ggml_mul_mat(ctx, A, B)`](https://github.com/ggerganov/llama.cpp/blob/880e352277fc017df4d5794f0c21c44e1eae2b84/ggml.h#L1058-L1064) means $C^T = A B^T \Leftrightarrow C = B A^T.$ + +![matmul](media/matmul.png) -* Please rate the complexity of your PR (i.e. `Review Complexity : Low`, `Review Complexity : Medium`, `Review Complexity : High`). This makes it easier for maintainers to triage the PRs. - - The PR template has a series of review complexity checkboxes `[ ]` that you can mark as `[X]` for your conveience. Refer to [About task lists](https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/about-task-lists) for more information. -* If the pull request only contains documentation changes (e.g., updating READMEs, adding new wiki pages), please add `[no ci]` to the commit title. This will skip unnecessary CI checks and help reduce build times. -* When squashing multiple commits on merge, use the following format for your commit title: ` : (#)`. For example: `utils : Fix typo in utils.py (#1234)` diff --git a/Makefile b/Makefile index bbfe0f12b..c82f4268a 100644 --- a/Makefile +++ b/Makefile @@ -11,9 +11,9 @@ BUILD_TARGETS = \ llama-embedding \ llama-eval-callback \ llama-export-lora \ - llama-finetune \ llama-gbnf-validator \ llama-gguf \ + llama-gguf-hash \ llama-gguf-split \ llama-gritlm \ llama-imatrix \ @@ -36,7 +36,6 @@ BUILD_TARGETS = \ llama-simple \ llama-speculative \ llama-tokenize \ - llama-train-text-from-scratch \ llama-vdot \ llama-cvector-generator \ tests/test-c.o @@ -45,6 +44,7 @@ BUILD_TARGETS = \ TEST_TARGETS = \ tests/test-autorelease \ tests/test-backend-ops \ + tests/test-chat-template \ tests/test-double-float \ tests/test-grad0 \ tests/test-grammar-integration \ @@ -61,6 +61,15 @@ TEST_TARGETS = \ tests/test-tokenizer-1-bpe \ tests/test-tokenizer-1-spm +# Legacy build targets that were renamed in #7809, but should still be removed when the project is cleaned +LEGACY_TARGETS_CLEAN = main quantize quantize-stats perplexity imatrix embedding vdot q8dot convert-llama2c-to-ggml \ + simple batched batched-bench save-load-state server gguf gguf-split eval-callback llama-bench libllava.a llava-cli baby-llama \ + retrieval speculative infill tokenize benchmark-matmult parallel export-lora lookahead lookup passkey gritlm + +# Legacy build targets that were renamed in #7809, but we want to build binaries that for them that output a deprecation warning if people try to use them. +# We don't want to clutter things too much, so we only build replacements for the most commonly used binaries. +LEGACY_TARGETS_BUILD = main quantize perplexity embedding server + # Deprecation aliases ifdef LLAMA_CUBLAS $(error LLAMA_CUBLAS is removed. Use GGML_CUDA instead.) @@ -186,7 +195,11 @@ ifdef GGML_RPC BUILD_TARGETS += rpc-server endif -default: $(BUILD_TARGETS) +ifdef GGML_VULKAN + BUILD_TARGETS += vulkan-shaders-gen +endif + +default: $(BUILD_TARGETS) $(LEGACY_TARGETS_BUILD) test: $(TEST_TARGETS) @failures=0; \ @@ -221,7 +234,7 @@ test: $(TEST_TARGETS) fi @echo 'All tests passed.' -all: $(BUILD_TARGETS) $(TEST_TARGETS) +all: $(BUILD_TARGETS) $(TEST_TARGETS) $(LEGACY_TARGETS_BUILD) ifdef RISCV_CROSS_COMPILE CC := riscv64-unknown-linux-gnu-gcc @@ -238,17 +251,22 @@ MK_CFLAGS = -std=c11 -fPIC MK_CXXFLAGS = -std=c++11 -fPIC MK_NVCCFLAGS = -std=c++11 -ifndef LLAMA_NO_CCACHE +ifdef LLAMA_NO_CCACHE +GGML_NO_CCACHE := 1 +DEPRECATE_WARNING := 1 +endif + +ifndef GGML_NO_CCACHE CCACHE := $(shell which ccache) ifdef CCACHE export CCACHE_SLOPPINESS = time_macros -$(info I ccache found, compilation results will be cached. Disable with LLAMA_NO_CCACHE.) +$(info I ccache found, compilation results will be cached. Disable with GGML_NO_CCACHE.) CC := $(CCACHE) $(CC) CXX := $(CCACHE) $(CXX) else $(info I ccache not found. Consider installing it for faster compilation.) endif # CCACHE -endif # LLAMA_NO_CCACHE +endif # GGML_NO_CCACHE # clock_gettime came in POSIX.1b (1993) # CLOCK_MONOTONIC came in POSIX.1-2001 / SUSv3 as optional @@ -307,9 +325,9 @@ ifdef LLAMA_DEBUG endif else MK_CPPFLAGS += -DNDEBUG - MK_CFLAGS += -O3 - MK_CXXFLAGS += -O3 - MK_NVCCFLAGS += -O3 + MK_CFLAGS += -O3 -g + MK_CXXFLAGS += -O3 -g + MK_NVCCFLAGS += -O3 -g endif ifdef LLAMA_SANITIZE_THREAD @@ -510,10 +528,21 @@ ifndef GGML_NO_ACCELERATE endif endif # GGML_NO_ACCELERATE +ifdef GGML_MUSA + CC := clang + CXX := clang++ + GGML_CUDA := 1 + MK_CPPFLAGS += -DGGML_USE_MUSA +endif + ifndef GGML_NO_OPENMP MK_CPPFLAGS += -DGGML_USE_OPENMP MK_CFLAGS += -fopenmp MK_CXXFLAGS += -fopenmp + ifdef GGML_MUSA + MK_CPPFLAGS += -I/usr/lib/llvm-10/include/openmp + MK_LDFLAGS += -L/usr/lib/llvm-10/lib + endif # GGML_MUSA endif # GGML_NO_OPENMP ifdef GGML_OPENBLAS @@ -531,14 +560,20 @@ ifdef GGML_OPENBLAS64 endif # GGML_OPENBLAS64 ifdef GGML_BLIS - MK_CPPFLAGS += -DGGML_USE_BLAS -I/usr/local/include/blis -I/usr/include/blis + MK_CPPFLAGS += -DGGML_USE_BLAS -DGGML_BLAS_USE_BLIS -I/usr/local/include/blis -I/usr/include/blis MK_LDFLAGS += -lblis -L/usr/local/lib OBJ_GGML += ggml/src/ggml-blas.o endif # GGML_BLIS +ifdef GGML_NVPL + MK_CPPFLAGS += -DGGML_USE_BLAS -DGGML_BLAS_USE_NVPL -DNVPL_ILP64 -I/usr/local/include/nvpl_blas -I/usr/include/nvpl_blas + MK_LDFLAGS += -L/usr/local/lib -lnvpl_blas_core -lnvpl_blas_ilp64_gomp + OBJ_GGML += ggml/src/ggml-blas.o +endif # GGML_NVPL + ifndef GGML_NO_LLAMAFILE MK_CPPFLAGS += -DGGML_USE_LLAMAFILE - OBJ_GGML += ggml/src/sgemm.o + OBJ_GGML += ggml/src/llamafile/sgemm.o endif ifdef GGML_RPC @@ -558,15 +593,27 @@ else endif # GGML_CUDA_FA_ALL_QUANTS ifdef GGML_CUDA - ifneq ('', '$(wildcard /opt/cuda)') - CUDA_PATH ?= /opt/cuda - else - CUDA_PATH ?= /usr/local/cuda - endif + ifdef GGML_MUSA + ifneq ('', '$(wildcard /opt/musa)') + CUDA_PATH ?= /opt/musa + else + CUDA_PATH ?= /usr/local/musa + endif - MK_CPPFLAGS += -DGGML_USE_CUDA -I$(CUDA_PATH)/include -I$(CUDA_PATH)/targets/$(UNAME_M)-linux/include -DGGML_CUDA_USE_GRAPHS - MK_LDFLAGS += -lcuda -lcublas -lculibos -lcudart -lcublasLt -lpthread -ldl -lrt -L$(CUDA_PATH)/lib64 -L/usr/lib64 -L$(CUDA_PATH)/targets/$(UNAME_M)-linux/lib -L$(CUDA_PATH)/lib64/stubs -L/usr/lib/wsl/lib - MK_NVCCFLAGS += -use_fast_math + MK_CPPFLAGS += -DGGML_USE_CUDA -I$(CUDA_PATH)/include + MK_LDFLAGS += -lmusa -lmublas -lmusart -lpthread -ldl -lrt -L$(CUDA_PATH)/lib -L/usr/lib64 + MK_NVCCFLAGS += -x musa -mtgpu --cuda-gpu-arch=mp_22 + else + ifneq ('', '$(wildcard /opt/cuda)') + CUDA_PATH ?= /opt/cuda + else + CUDA_PATH ?= /usr/local/cuda + endif + + MK_CPPFLAGS += -DGGML_USE_CUDA -I$(CUDA_PATH)/include -I$(CUDA_PATH)/targets/$(UNAME_M)-linux/include -DGGML_CUDA_USE_GRAPHS + MK_LDFLAGS += -lcuda -lcublas -lculibos -lcudart -lcublasLt -lpthread -ldl -lrt -L$(CUDA_PATH)/lib64 -L/usr/lib64 -L$(CUDA_PATH)/targets/$(UNAME_M)-linux/lib -L$(CUDA_PATH)/lib64/stubs -L/usr/lib/wsl/lib + MK_NVCCFLAGS += -use_fast_math + endif # GGML_MUSA OBJ_GGML += ggml/src/ggml-cuda.o OBJ_GGML += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/*.cu)) @@ -576,9 +623,11 @@ ifdef LLAMA_FATAL_WARNINGS MK_NVCCFLAGS += -Werror all-warnings endif # LLAMA_FATAL_WARNINGS +ifndef GGML_MUSA ifndef JETSON_EOL_MODULE_DETECT MK_NVCCFLAGS += --forward-unknown-to-host-compiler endif # JETSON_EOL_MODULE_DETECT +endif # GGML_MUSA ifdef LLAMA_DEBUG MK_NVCCFLAGS += -lineinfo @@ -591,8 +640,12 @@ endif # GGML_CUDA_DEBUG ifdef GGML_CUDA_NVCC NVCC = $(CCACHE) $(GGML_CUDA_NVCC) else - NVCC = $(CCACHE) nvcc -endif #GGML_CUDA_NVCC + ifdef GGML_MUSA + NVCC = $(CCACHE) mcc + else + NVCC = $(CCACHE) nvcc + endif # GGML_MUSA +endif # GGML_CUDA_NVCC ifdef CUDA_DOCKER_ARCH MK_NVCCFLAGS += -Wno-deprecated-gpu-targets -arch=$(CUDA_DOCKER_ARCH) @@ -663,9 +716,15 @@ define NVCC_COMPILE $(NVCC) -I. -Icommon -D_XOPEN_SOURCE=600 -D_GNU_SOURCE -DNDEBUG -DGGML_USE_CUDA -I/usr/local/cuda/include -I/opt/cuda/include -I/usr/local/cuda/targets/aarch64-linux/include -std=c++11 -O3 $(NVCCFLAGS) $(CPPFLAGS) -Xcompiler "$(CUDA_CXXFLAGS)" -c $< -o $@ endef # NVCC_COMPILE else + ifdef GGML_MUSA +define NVCC_COMPILE + $(NVCC) $(NVCCFLAGS) $(CPPFLAGS) -c $< -o $@ +endef # NVCC_COMPILE + else define NVCC_COMPILE $(NVCC) $(NVCCFLAGS) $(CPPFLAGS) -Xcompiler "$(CUDA_CXXFLAGS)" -c $< -o $@ endef # NVCC_COMPILE + endif # GGML_MUSA endif # JETSON_EOL_MODULE_DETECT ggml/src/ggml-cuda/%.o: \ @@ -688,8 +747,8 @@ endif # GGML_CUDA ifdef GGML_VULKAN MK_CPPFLAGS += -DGGML_USE_VULKAN - MK_LDFLAGS += -lvulkan - OBJ_GGML += ggml/src/ggml-vulkan.o + MK_LDFLAGS += $(shell pkg-config --libs vulkan) + OBJ_GGML += ggml/src/ggml-vulkan.o ggml/src/ggml-vulkan-shaders.o ifdef GGML_VULKAN_CHECK_RESULTS MK_CPPFLAGS += -DGGML_VULKAN_CHECK_RESULTS @@ -711,10 +770,28 @@ ifdef GGML_VULKAN_RUN_TESTS MK_CPPFLAGS += -DGGML_VULKAN_RUN_TESTS endif -ggml/src/ggml-vulkan.o: \ - ggml/src/ggml-vulkan.cpp \ - ggml/include/ggml-vulkan.h - $(CXX) $(CXXFLAGS) -c $< -o $@ +GLSLC_CMD = glslc +_ggml_vk_genshaders_cmd = $(shell pwd)/vulkan-shaders-gen +_ggml_vk_header = ggml/src/ggml-vulkan-shaders.hpp +_ggml_vk_source = ggml/src/ggml-vulkan-shaders.cpp +_ggml_vk_input_dir = ggml/src/vulkan-shaders +_ggml_vk_shader_deps = $(echo $(_ggml_vk_input_dir)/*.comp) + +ggml/src/ggml-vulkan.o: ggml/src/ggml-vulkan.cpp ggml/include/ggml-vulkan.h $(_ggml_vk_header) $(_ggml_vk_source) + $(CXX) $(CXXFLAGS) $(shell pkg-config --cflags vulkan) -c $< -o $@ + +$(_ggml_vk_header): $(_ggml_vk_source) + +$(_ggml_vk_source): $(_ggml_vk_shader_deps) vulkan-shaders-gen + $(_ggml_vk_genshaders_cmd) \ + --glslc $(GLSLC_CMD) \ + --input-dir $(_ggml_vk_input_dir) \ + --target-hpp $(_ggml_vk_header) \ + --target-cpp $(_ggml_vk_source) + +vulkan-shaders-gen: ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp + $(CXX) $(CXXFLAGS) -o $@ $(LDFLAGS) ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp + endif # GGML_VULKAN ifdef GGML_HIPBLAS @@ -751,6 +828,14 @@ ifdef GGML_CUDA_FORCE_DMMV HIPFLAGS += -DGGML_CUDA_FORCE_DMMV endif # GGML_CUDA_FORCE_DMMV +ifdef GGML_CUDA_FORCE_MMQ + HIPFLAGS += -DGGML_CUDA_FORCE_MMQ +endif # GGML_CUDA_FORCE_MMQ + +ifdef GGML_CUDA_FORCE_CUBLAS + HIPFLAGS += -DGGML_CUDA_FORCE_CUBLAS +endif # GGML_CUDA_FORCE_CUBLAS + ifdef GGML_CUDA_NO_PEER_COPY HIPFLAGS += -DGGML_CUDA_NO_PEER_COPY endif # GGML_CUDA_NO_PEER_COPY @@ -819,10 +904,14 @@ OBJ_GGML += \ ggml/src/ggml.o \ ggml/src/ggml-alloc.o \ ggml/src/ggml-backend.o \ - ggml/src/ggml-quants.o + ggml/src/ggml-quants.o \ + ggml/src/ggml-aarch64.o OBJ_LLAMA = \ src/llama.o \ + src/llama-vocab.o \ + src/llama-grammar.o \ + src/llama-sampling.o \ src/unicode.o \ src/unicode-data.o @@ -890,6 +979,7 @@ $(info I CXX: $(shell $(CXX) --version | head -n 1)) ifdef GGML_CUDA $(info I NVCC: $(shell $(NVCC) --version | tail -n 1)) CUDA_VERSION := $(shell $(NVCC) --version | grep -oP 'release (\K[0-9]+\.[0-9])') +ifndef GGML_MUSA ifeq ($(shell awk -v "v=$(CUDA_VERSION)" 'BEGIN { print (v < 11.7) }'),1) ifndef CUDA_DOCKER_ARCH @@ -899,6 +989,7 @@ endif # CUDA_POWER_ARCH endif # CUDA_DOCKER_ARCH endif # eq ($(shell echo "$(CUDA_VERSION) < 11.7" | bc),1) +endif # GGML_MUSA endif # GGML_CUDA $(info ) @@ -919,6 +1010,7 @@ $(info - LLAMA_NO_LLAMAFILE) $(info - LLAMA_NO_ACCELERATE) $(info - LLAMA_NO_OPENMP) $(info - LLAMA_NO_METAL) +$(info - LLAMA_NO_CCACHE) $(info ) endif @@ -952,15 +1044,22 @@ ggml/src/ggml-quants.o: \ ggml/src/ggml-common.h $(CC) $(CFLAGS) -c $< -o $@ +ggml/src/ggml-aarch64.o: \ + ggml/src/ggml-aarch64.c \ + ggml/include/ggml.h \ + ggml/src/ggml-aarch64.h \ + ggml/src/ggml-common.h + $(CC) $(CFLAGS) -c $< -o $@ + ggml/src/ggml-blas.o: \ ggml/src/ggml-blas.cpp \ ggml/include/ggml-blas.h $(CXX) $(CXXFLAGS) -c $< -o $@ ifndef GGML_NO_LLAMAFILE -ggml/src/sgemm.o: \ - ggml/src/sgemm.cpp \ - ggml/src/sgemm.h \ +ggml/src/llamafile/sgemm.o: \ + ggml/src/llamafile/sgemm.cpp \ + ggml/src/llamafile/sgemm.h \ ggml/include/ggml.h $(CXX) $(CXXFLAGS) -c $< -o $@ endif # GGML_NO_LLAMAFILE @@ -994,6 +1093,10 @@ src/unicode-data.o: \ src/llama.o: \ src/llama.cpp \ + src/llama-impl.h \ + src/llama-vocab.h \ + src/llama-grammar.h \ + src/llama-sampling.h \ src/unicode.h \ include/llama.h \ ggml/include/ggml-cuda.h \ @@ -1003,6 +1106,29 @@ src/llama.o: \ ggml/include/ggml-backend.h $(CXX) $(CXXFLAGS) -c $< -o $@ +src/llama-vocab.o: \ + src/llama-vocab.cpp \ + src/llama-vocab.h \ + src/llama-impl.h \ + include/llama.h + $(CXX) $(CXXFLAGS) -c $< -o $@ + +src/llama-grammar.o: \ + src/llama-grammar.cpp \ + src/llama-grammar.h \ + src/llama-impl.h \ + src/llama-vocab.h \ + src/llama-sampling.h \ + include/llama.h + $(CXX) $(CXXFLAGS) -c $< -o $@ + +src/llama-sampling.o: \ + src/llama-sampling.cpp \ + src/llama-sampling.h \ + src/llama-impl.h \ + include/llama.h + $(CXX) $(CXXFLAGS) -c $< -o $@ + $(LIB_LLAMA): \ $(OBJ_LLAMA) \ $(LIB_GGML) @@ -1070,6 +1196,7 @@ clean: rm -rvf src/*.o rm -rvf tests/*.o rm -rvf examples/*.o + rm -rvf common/*.o rm -rvf *.a rm -rvf *.dll rm -rvf *.so @@ -1084,6 +1211,8 @@ clean: rm -vrf ggml/src/ggml-cuda/template-instances/*.o rm -rvf $(BUILD_TARGETS) rm -rvf $(TEST_TARGETS) + rm -f vulkan-shaders-gen ggml/src/ggml-vulkan-shaders.hpp ggml/src/ggml-vulkan-shaders.cpp + rm -rvf $(LEGACY_TARGETS_CLEAN) find examples pocs -type f -name "*.o" -delete # @@ -1170,6 +1299,23 @@ llama-gguf: examples/gguf/gguf.cpp \ $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) +examples/gguf-hash/deps/sha1/sha1.o: \ + examples/gguf-hash/deps/sha1/sha1.c + $(CC) $(CFLAGS) -Iexamples/gguf-hash/deps -c $< -o $@ + +examples/gguf-hash/deps/xxhash/xxhash.o: \ + examples/gguf-hash/deps/xxhash/xxhash.c + $(CC) $(CFLAGS) -Iexamples/gguf-hash/deps -c $< -o $@ + +examples/gguf-hash/deps/sha256/sha256.o: \ + examples/gguf-hash/deps/sha256/sha256.c + $(CC) $(CFLAGS) -Iexamples/gguf-hash/deps -c $< -o $@ + +llama-gguf-hash: examples/gguf-hash/gguf-hash.cpp examples/gguf-hash/deps/sha1/sha1.o examples/gguf-hash/deps/xxhash/xxhash.o examples/gguf-hash/deps/sha256/sha256.o\ + $(OBJ_ALL) + $(CXX) $(CXXFLAGS) -Iexamples/gguf-hash/deps -c $< -o $(call GET_OBJ_FILE, $<) + $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) + llama-gguf-split: examples/gguf-split/gguf-split.cpp \ $(OBJ_ALL) $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) @@ -1185,11 +1331,6 @@ llama-cvector-generator: examples/cvector-generator/cvector-generator.cpp \ $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) -llama-train-text-from-scratch: examples/train-text-from-scratch/train-text-from-scratch.cpp \ - $(OBJ_ALL) - $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) - $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) - llama-convert-llama2c-to-ggml: examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp \ $(OBJ_GGML) $(OBJ_LLAMA) $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) @@ -1205,13 +1346,8 @@ llama-baby-llama: examples/baby-llama/baby-llama.cpp \ $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) -llama-finetune: examples/finetune/finetune.cpp \ - $(OBJ_ALL) - $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) - $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) - llama-export-lora: examples/export-lora/export-lora.cpp \ - $(OBJ_GGML) common/log.h + $(OBJ_ALL) $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) @@ -1358,7 +1494,7 @@ run-benchmark-matmult: llama-benchmark-matmult .PHONY: run-benchmark-matmult swift tests/test-llama-grammar: tests/test-llama-grammar.cpp \ - $(OBJ_GGML) $(OBJ_COMMON) src/unicode.o src/unicode-data.o + $(OBJ_ALL) $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) @@ -1462,3 +1598,51 @@ llama-q8dot: pocs/vdot/q8dot.cpp ggml/src/ggml.o \ $(OBJ_GGML) $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) $(CXX) $(CXXFLAGS) $(filter-out $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) + +# +# Deprecated binaries that we want to keep around long enough for people to migrate to the new filenames, then these can be removed. +# +# Mark legacy binary targets as .PHONY so that they are always checked. +.PHONY: main quantize perplexity embedding server + +# NOTE: We currently will always build the deprecation-warning `main` and `server` binaries to help users migrate. +# Eventually we will want to remove these target from building all the time. +main: examples/deprecation-warning/deprecation-warning.cpp + $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) + $(CXX) $(CXXFLAGS) $(filter-out $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) + @echo "NOTICE: The 'main' binary is deprecated. Please use 'llama-cli' instead." + +server: examples/deprecation-warning/deprecation-warning.cpp + $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) + $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) + @echo "NOTICE: The 'server' binary is deprecated. Please use 'llama-server' instead." + +quantize: examples/deprecation-warning/deprecation-warning.cpp +ifneq (,$(wildcard quantize)) + $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) + $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) + @echo "#########" + @echo "WARNING: The 'quantize' binary is deprecated. Please use 'llama-quantize' instead." + @echo " Remove the 'quantize' binary to remove this warning." + @echo "#########" +endif + +perplexity: examples/deprecation-warning/deprecation-warning.cpp +ifneq (,$(wildcard perplexity)) + $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) + $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) + @echo "#########" + @echo "WARNING: The 'perplexity' binary is deprecated. Please use 'llama-perplexity' instead." + @echo " Remove the 'perplexity' binary to remove this warning." + @echo "#########" +endif + +embedding: examples/deprecation-warning/deprecation-warning.cpp +ifneq (,$(wildcard embedding)) + $(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<) + $(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) + @echo "#########" + @echo "WARNING: The 'embedding' binary is deprecated. Please use 'llama-embedding' instead." + @echo " Remove the 'embedding' binary to remove this warning." + @echo "#########" +endif diff --git a/Package.swift b/Package.swift index 77fed86df..1d90b47bf 100644 --- a/Package.swift +++ b/Package.swift @@ -4,12 +4,16 @@ import PackageDescription var sources = [ "src/llama.cpp", + "src/llama-vocab.cpp", + "src/llama-grammar.cpp", + "src/llama-sampling.cpp", "src/unicode.cpp", "src/unicode-data.cpp", "ggml/src/ggml.c", "ggml/src/ggml-alloc.c", "ggml/src/ggml-backend.c", "ggml/src/ggml-quants.c", + "ggml/src/ggml-aarch64.c", ] var resources: [Resource] = [] diff --git a/README.md b/README.md index 99b16f6e2..775ce2c88 100644 --- a/README.md +++ b/README.md @@ -3,7 +3,7 @@ ![llama](https://user-images.githubusercontent.com/1991296/230134379-7181e485-c521-4d23-a0d6-f7b3b61ba524.png) [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](https://opensource.org/licenses/MIT) -[![Server](https://github.com/ggerganov/llama.cpp/actions/workflows/server.yml/badge.svg?branch=master&event=schedule)](https://github.com/ggerganov/llama.cpp/actions/workflows/server.yml) +[![Server](https://github.com/ggerganov/llama.cpp/actions/workflows/server.yml/badge.svg)](https://github.com/ggerganov/llama.cpp/actions/workflows/server.yml) [![Conan Center](https://shields.io/conan/v/llama-cpp)](https://conan.io/center/llama-cpp) [Roadmap](https://github.com/users/ggerganov/projects/7) / [Project status](https://github.com/ggerganov/llama.cpp/discussions/3471) / [Manifesto](https://github.com/ggerganov/llama.cpp/discussions/205) / [ggml](https://github.com/ggerganov/ggml) @@ -13,7 +13,7 @@ Inference of Meta's [LLaMA](https://arxiv.org/abs/2302.13971) model (and others) > [!IMPORTANT] [2024 Jun 12] Binaries have been renamed w/ a `llama-` prefix. `main` is now `llama-cli`, `server` is `llama-server`, etc (https://github.com/ggerganov/llama.cpp/pull/7809) -### Recent API changes +## Recent API changes - [2024 Jun 26] The source code and CMake build scripts have been restructured https://github.com/ggerganov/llama.cpp/pull/8006 - [2024 Apr 21] `llama_token_to_piece` can now optionally render special tokens https://github.com/ggerganov/llama.cpp/pull/6807 @@ -24,9 +24,9 @@ Inference of Meta's [LLaMA](https://arxiv.org/abs/2302.13971) model (and others) - [2024 Mar 4] Embeddings API updated https://github.com/ggerganov/llama.cpp/pull/5796 - [2024 Mar 3] `struct llama_context_params` https://github.com/ggerganov/llama.cpp/pull/5849 -### Hot topics +## Hot topics -- **`convert.py` has been deprecated and moved to `examples/convert-legacy-llama.py`, please use `convert-hf-to-gguf.py`** https://github.com/ggerganov/llama.cpp/pull/7430 +- **`convert.py` has been deprecated and moved to `examples/convert_legacy_llama.py`, please use `convert_hf_to_gguf.py`** https://github.com/ggerganov/llama.cpp/pull/7430 - Initial Flash-Attention support: https://github.com/ggerganov/llama.cpp/pull/5021 - BPE pre-tokenization support has been added: https://github.com/ggerganov/llama.cpp/pull/6920 - MoE memory layout has been updated - reconvert models for `mmap` support and regenerate `imatrix` https://github.com/ggerganov/llama.cpp/pull/6387 @@ -39,37 +39,6 @@ Inference of Meta's [LLaMA](https://arxiv.org/abs/2302.13971) model (and others) ---- -
- Table of Contents -
    -
  1. - Description -
    2. -
  2. - Usage - -
    3. -
  3. Contributing
    4. -
  4. Coding guidelines
    5. -
  5. Docs
    6. -
-
- ## Description The main goal of `llama.cpp` is to enable LLM inference with minimal setup and state-of-the-art performance on a wide @@ -87,14 +56,6 @@ Since its [inception](https://github.com/ggerganov/llama.cpp/issues/33#issuecomm improved significantly thanks to many contributions. It is the main playground for developing new features for the [ggml](https://github.com/ggerganov/ggml) library. -**Supported platforms:** - -- [X] Mac OS -- [X] Linux -- [X] Windows (via CMake) -- [X] Docker -- [X] FreeBSD - **Supported models:** Typically finetunes of the base models below are supported as well. @@ -108,6 +69,7 @@ Typically finetunes of the base models below are supported as well. - [X] [Falcon](https://huggingface.co/models?search=tiiuae/falcon) - [X] [Chinese LLaMA / Alpaca](https://github.com/ymcui/Chinese-LLaMA-Alpaca) and [Chinese LLaMA-2 / Alpaca-2](https://github.com/ymcui/Chinese-LLaMA-Alpaca-2) - [X] [Vigogne (French)](https://github.com/bofenghuang/vigogne) +- [X] [BERT](https://github.com/ggerganov/llama.cpp/pull/5423) - [X] [Koala](https://bair.berkeley.edu/blog/2023/04/03/koala/) - [X] [Baichuan 1 & 2](https://huggingface.co/models?search=baichuan-inc/Baichuan) + [derivations](https://huggingface.co/hiyouga/baichuan-7b-sft) - [X] [Aquila 1 & 2](https://huggingface.co/models?search=BAAI/Aquila) @@ -134,8 +96,9 @@ Typically finetunes of the base models below are supported as well. - [x] [GritLM-7B](https://huggingface.co/GritLM/GritLM-7B) + [GritLM-8x7B](https://huggingface.co/GritLM/GritLM-8x7B) - [x] [OLMo](https://allenai.org/olmo) - [x] [GPT-NeoX](https://github.com/EleutherAI/gpt-neox) + [Pythia](https://github.com/EleutherAI/pythia) +- [x] [ChatGLM3-6b](https://huggingface.co/THUDM/chatglm3-6b) + [ChatGLM4-9b](https://huggingface.co/THUDM/glm-4-9b) -(instructions for supporting more models: [HOWTO-add-model.md](./docs/HOWTO-add-model.md)) +(instructions for supporting more models: [HOWTO-add-model.md](./docs/development/HOWTO-add-model.md)) **Multimodal models:** @@ -149,12 +112,6 @@ Typically finetunes of the base models below are supported as well. - [x] [Moondream](https://huggingface.co/vikhyatk/moondream2) - [x] [Bunny](https://github.com/BAAI-DCAI/Bunny) -**HTTP server** - -[llama.cpp web server](./examples/server) is a lightweight [OpenAI API](https://github.com/openai/openai-openapi) compatible HTTP server that can be used to serve local models and easily connect them to existing clients. - -[simplechat](./examples/server/public_simplechat) is a simple chat client, which can be used to chat with the model exposed using above web server (use --path to point to simplechat), from a local web browser. - **Bindings:** - Python: [abetlen/llama-cpp-python](https://github.com/abetlen/llama-cpp-python) @@ -175,11 +132,13 @@ Typically finetunes of the base models below are supported as well. - Zig: [deins/llama.cpp.zig](https://github.com/Deins/llama.cpp.zig) - Flutter/Dart: [netdur/llama_cpp_dart](https://github.com/netdur/llama_cpp_dart) - PHP (API bindings and features built on top of llama.cpp): [distantmagic/resonance](https://github.com/distantmagic/resonance) [(more info)](https://github.com/ggerganov/llama.cpp/pull/6326) +- Guile Scheme: [guile_llama_cpp](https://savannah.nongnu.org/projects/guile-llama-cpp) **UI:** Unless otherwise noted these projects are open-source with permissive licensing: +- [MindWorkAI/AI-Studio](https://github.com/MindWorkAI/AI-Studio) (FSL-1.1-MIT) - [iohub/collama](https://github.com/iohub/coLLaMA) - [janhq/jan](https://github.com/janhq/jan) (AGPL) - [nat/openplayground](https://github.com/nat/openplayground) @@ -217,10 +176,19 @@ Unless otherwise noted these projects are open-source with permissive licensing: **Tools:** - [akx/ggify](https://github.com/akx/ggify) – download PyTorch models from HuggingFace Hub and convert them to GGML +- [crashr/gppm](https://github.com/crashr/gppm) – launch llama.cpp instances utilizing NVIDIA Tesla P40 or P100 GPUs with reduced idle power consumption ---- +**Infrastructure:** -Here is a typical run using LLaMA v2 13B on M2 Ultra: +- [Paddler](https://github.com/distantmagic/paddler) - Stateful load balancer custom-tailored for llama.cpp + +**Games:** +- [Lucy's Labyrinth](https://github.com/MorganRO8/Lucys_Labyrinth) - A simple maze game where agents controlled by an AI model will try to trick you. + +## Demo + +
+Typical run using LLaMA v2 13B on M2 Ultra ``` $ make -j && ./llama-cli -m models/llama-13b-v2/ggml-model-q4_0.gguf -p "Building a website can be done in 10 simple steps:\nStep 1:" -n 400 -e @@ -300,454 +268,85 @@ llama_print_timings: eval time = 24513.59 ms / 399 runs ( 61.44 ms llama_print_timings: total time = 25431.49 ms ``` +
+ +
+Demo of running both LLaMA-7B and whisper.cpp on a single M1 Pro MacBook + And here is another demo of running both LLaMA-7B and [whisper.cpp](https://github.com/ggerganov/whisper.cpp) on a single M1 Pro MacBook: https://user-images.githubusercontent.com/1991296/224442907-7693d4be-acaa-4e01-8b4f-add84093ffff.mp4 +
+ ## Usage Here are the end-to-end binary build and model conversion steps for most supported models. -### Get the Code +### Basic usage + +Firstly, you need to get the binary. There are different methods that you can follow: +- Method 1: Clone this repository and build locally, see [how to build](./docs/build.md) +- Method 2: If you are using MacOS or Linux, you can install llama.cpp via [brew, flox or nix](./docs/install.md) +- Method 3: Use a Docker image, see [documentation for Docker](./docs/docker.md) +- Method 4: Download pre-built binary from [releases](https://github.com/ggerganov/llama.cpp/releases) + +You can run a basic completion using this command: ```bash -git clone https://github.com/ggerganov/llama.cpp -cd llama.cpp +llama-cli -m your_model.gguf -p "I believe the meaning of life is" -n 128 + +# Output: +# I believe the meaning of life is to find your own truth and to live in accordance with it. For me, this means being true to myself and following my passions, even if they don't align with societal expectations. I think that's what I love about yoga – it's not just a physical practice, but a spiritual one too. It's about connecting with yourself, listening to your inner voice, and honoring your own unique journey. ``` -### Build +See [this page](./examples/main/README.md) for a full list of parameters. -In order to build llama.cpp you have four different options. +### Conversation mode -- Using `make`: - - On Linux or MacOS: - - ```bash - make - ``` - - - On Windows: - - 1. Download the latest fortran version of [w64devkit](https://github.com/skeeto/w64devkit/releases). - 2. Extract `w64devkit` on your pc. - 3. Run `w64devkit.exe`. - 4. Use the `cd` command to reach the `llama.cpp` folder. - 5. From here you can run: - ```bash - make - ``` - - - Notes: - - For faster compilation, add the `-j` argument to run multiple jobs in parallel. For example, `make -j 8` will run 8 jobs in parallel. - - For faster repeated compilation, install [ccache](https://ccache.dev/). - - For debug builds, run `make LLAMA_DEBUG=1` - -- Using `CMake`: - - ```bash - cmake -B build - cmake --build build --config Release - ``` - - **Notes**: - - - For faster compilation, add the `-j` argument to run multiple jobs in parallel. For example, `cmake --build build --config Release -j 8` will run 8 jobs in parallel. - - For faster repeated compilation, install [ccache](https://ccache.dev/). - - For debug builds, there are two cases: - - 1. Single-config generators (e.g. default = `Unix Makefiles`; note that they just ignore the `--config` flag): - - ```bash - cmake -B build -DCMAKE_BUILD_TYPE=Debug - cmake --build build - ``` - - 2. Multi-config generators (`-G` param set to Visual Studio, XCode...): - - ```bash - cmake -B build -G "Xcode" - cmake --build build --config Debug - ``` - -- Using `gmake` (FreeBSD): - - 1. Install and activate [DRM in FreeBSD](https://wiki.freebsd.org/Graphics) - 2. Add your user to **video** group - 3. Install compilation dependencies. - - ```bash - sudo pkg install gmake automake autoconf pkgconf llvm15 openblas - - gmake CC=/usr/local/bin/clang15 CXX=/usr/local/bin/clang++15 -j4 - ``` - -### Homebrew - -On Mac and Linux, the homebrew package manager can be used via -``` -brew install llama.cpp -``` -The formula is automatically updated with new `llama.cpp` releases. More info: https://github.com/ggerganov/llama.cpp/discussions/7668 - -### Nix - -On Mac and Linux, the Nix package manager can be used via -``` -nix profile install nixpkgs#llama-cpp -``` -For flake enabled installs. - -Or -``` -nix-env --file '' --install --attr llama-cpp -``` -For non-flake enabled installs. - -This expression is automatically updated within the [nixpkgs repo](https://github.com/NixOS/nixpkgs/blob/nixos-24.05/pkgs/by-name/ll/llama-cpp/package.nix#L164). - -#### Flox - -On Mac and Linux, Flox can be used to install llama.cpp within a Flox environment via -``` -flox install llama-cpp -``` -Flox follows the nixpkgs build of llama.cpp. - -### Metal Build - -On MacOS, Metal is enabled by default. Using Metal makes the computation run on the GPU. -To disable the Metal build at compile time use the `GGML_NO_METAL=1` flag or the `GGML_METAL=OFF` cmake option. - -When built with Metal support, you can explicitly disable GPU inference with the `--n-gpu-layers|-ngl 0` command-line -argument. - -### BLAS Build - -Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). Support with CPU-only BLAS implementations doesn't affect the normal generation performance. We may see generation performance improvements with GPU-involved BLAS implementations, e.g. cuBLAS, hipBLAS. There are currently several different BLAS implementations available for build and use: - -- #### Accelerate Framework: - - This is only available on Mac PCs and it's enabled by default. You can just build using the normal instructions. - -- #### OpenBLAS: - - This provides BLAS acceleration using only the CPU. Make sure to have OpenBLAS installed on your machine. - - - Using `make`: - - On Linux: - ```bash - make GGML_OPENBLAS=1 - ``` - - - On Windows: - - 1. Download the latest fortran version of [w64devkit](https://github.com/skeeto/w64devkit/releases). - 2. Download the latest version of [OpenBLAS for Windows](https://github.com/xianyi/OpenBLAS/releases). - 3. Extract `w64devkit` on your pc. - 4. From the OpenBLAS zip that you just downloaded copy `libopenblas.a`, located inside the `lib` folder, inside `w64devkit\x86_64-w64-mingw32\lib`. - 5. From the same OpenBLAS zip copy the content of the `include` folder inside `w64devkit\x86_64-w64-mingw32\include`. - 6. Run `w64devkit.exe`. - 7. Use the `cd` command to reach the `llama.cpp` folder. - 8. From here you can run: - - ```bash - make GGML_OPENBLAS=1 - ``` - - - Using `CMake` on Linux: - - ```bash - cmake -B build -DGGML_BLAS=ON -DGGML_BLAS_VENDOR=OpenBLAS - cmake --build build --config Release - ``` - -- #### BLIS - - Check [BLIS.md](docs/BLIS.md) for more information. - -- #### SYCL - SYCL is a higher-level programming model to improve programming productivity on various hardware accelerators. - - llama.cpp based on SYCL is used to **support Intel GPU** (Data Center Max series, Flex series, Arc series, Built-in GPU and iGPU). - - For detailed info, please refer to [llama.cpp for SYCL](README-sycl.md). - -- #### Intel oneMKL - Building through oneAPI compilers will make avx_vnni instruction set available for intel processors that do not support avx512 and avx512_vnni. Please note that this build config **does not support Intel GPU**. For Intel GPU support, please refer to [llama.cpp for SYCL](./README-sycl.md). - - - Using manual oneAPI installation: - By default, `GGML_BLAS_VENDOR` is set to `Generic`, so if you already sourced intel environment script and assign `-DGGML_BLAS=ON` in cmake, the mkl version of Blas will automatically been selected. Otherwise please install oneAPI and follow the below steps: - ```bash - source /opt/intel/oneapi/setvars.sh # You can skip this step if in oneapi-basekit docker image, only required for manual installation - cmake -B build -DGGML_BLAS=ON -DGGML_BLAS_VENDOR=Intel10_64lp -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_NATIVE=ON - cmake --build build --config Release - ``` - - - Using oneAPI docker image: - If you do not want to source the environment vars and install oneAPI manually, you can also build the code using intel docker container: [oneAPI-basekit](https://hub.docker.com/r/intel/oneapi-basekit). Then, you can use the commands given above. - - Check [Optimizing and Running LLaMA2 on Intel® CPU](https://www.intel.com/content/www/us/en/content-details/791610/optimizing-and-running-llama2-on-intel-cpu.html) for more information. - -- #### CUDA - - This provides GPU acceleration using the CUDA cores of your Nvidia GPU. Make sure to have the CUDA toolkit installed. You can download it from your Linux distro's package manager (e.g. `apt install nvidia-cuda-toolkit`) or from here: [CUDA Toolkit](https://developer.nvidia.com/cuda-downloads). - - For Jetson user, if you have Jetson Orin, you can try this: [Offical Support](https://www.jetson-ai-lab.com/tutorial_text-generation.html). If you are using an old model(nano/TX2), need some additional operations before compiling. - - - Using `make`: - ```bash - make GGML_CUDA=1 - ``` - - Using `CMake`: - - ```bash - cmake -B build -DGGML_CUDA=ON - cmake --build build --config Release - ``` - - The environment variable [`CUDA_VISIBLE_DEVICES`](https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars) can be used to specify which GPU(s) will be used. The following compilation options are also available to tweak performance: - - | Option | Legal values | Default | Description | - |-------------------------------|------------------------|---------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| - | GGML_CUDA_FORCE_DMMV | Boolean | false | Force the use of dequantization + matrix vector multiplication kernels instead of using kernels that do matrix vector multiplication on quantized data. By default the decision is made based on compute capability (MMVQ for 6.1/Pascal/GTX 1000 or higher). Does not affect k-quants. | - | GGML_CUDA_DMMV_X | Positive integer >= 32 | 32 | Number of values in x direction processed by the CUDA dequantization + matrix vector multiplication kernel per iteration. Increasing this value can improve performance on fast GPUs. Power of 2 heavily recommended. Does not affect k-quants. | - | GGML_CUDA_MMV_Y | Positive integer | 1 | Block size in y direction for the CUDA mul mat vec kernels. Increasing this value can improve performance on fast GPUs. Power of 2 recommended. | - | GGML_CUDA_FORCE_MMQ | Boolean | false | Force the use of custom matrix multiplication kernels for quantized models instead of FP16 cuBLAS even if there is no int8 tensor core implementation available (affects V100, RDNA3). MMQ kernels are enabled by default on GPUs with int8 tensor core support. With MMQ force enabled, speed for large batch sizes will be worse but VRAM consumption will be lower. | - | GGML_CUDA_FORCE_CUBLAS | Boolean | false | Force the use of FP16 cuBLAS instead of custom matrix multiplication kernels for quantized models | - | GGML_CUDA_F16 | Boolean | false | If enabled, use half-precision floating point arithmetic for the CUDA dequantization + mul mat vec kernels and for the q4_1 and q5_1 matrix matrix multiplication kernels. Can improve performance on relatively recent GPUs. | - | GGML_CUDA_KQUANTS_ITER | 1 or 2 | 2 | Number of values processed per iteration and per CUDA thread for Q2_K and Q6_K quantization formats. Setting this value to 1 can improve performance for slow GPUs. | - | GGML_CUDA_PEER_MAX_BATCH_SIZE | Positive integer | 128 | Maximum batch size for which to enable peer access between multiple GPUs. Peer access requires either Linux or NVLink. When using NVLink enabling peer access for larger batch sizes is potentially beneficial. | - | GGML_CUDA_FA_ALL_QUANTS | Boolean | false | Compile support for all KV cache quantization type (combinations) for the FlashAttention CUDA kernels. More fine-grained control over KV cache size but compilation takes much longer. | - -- #### hipBLAS - - This provides BLAS acceleration on HIP-supported AMD GPUs. - Make sure to have ROCm installed. - You can download it from your Linux distro's package manager or from here: [ROCm Quick Start (Linux)](https://rocm.docs.amd.com/projects/install-on-linux/en/latest/tutorial/quick-start.html#rocm-install-quick). - - - Using `make`: - ```bash - make GGML_HIPBLAS=1 - ``` - - Using `CMake` for Linux (assuming a gfx1030-compatible AMD GPU): - ```bash - HIPCXX="$(hipconfig -l)/clang" HIP_PATH="$(hipconfig -R)" \ - cmake -S . -B build -DGGML_HIPBLAS=ON -DAMDGPU_TARGETS=gfx1030 -DCMAKE_BUILD_TYPE=Release \ - && cmake --build build --config Release -- -j 16 - ``` - On Linux it is also possible to use unified memory architecture (UMA) to share main memory between the CPU and integrated GPU by setting `-DGGML_HIP_UMA=ON`. - However, this hurts performance for non-integrated GPUs (but enables working with integrated GPUs). - - Note that if you get the following error: - ``` - clang: error: cannot find ROCm device library; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library - ``` - Try searching for a directory under `HIP_PATH` that contains the file - `oclc_abi_version_400.bc`. Then, add the following to the start of the - command: `HIP_DEVICE_LIB_PATH=`, so something - like: - ```bash - HIPCXX="$(hipconfig -l)/clang" HIP_PATH="$(hipconfig -p)" \ - HIP_DEVICE_LIB_PATH= \ - cmake -S . -B build -DGGML_HIPBLAS=ON -DAMDGPU_TARGETS=gfx1030 -DCMAKE_BUILD_TYPE=Release \ - && cmake --build build -- -j 16 - ``` - - - Using `make` (example for target gfx1030, build with 16 CPU threads): - ```bash - make -j16 GGML_HIPBLAS=1 GGML_HIP_UMA=1 AMDGPU_TARGETS=gfx1030 - ``` - - - Using `CMake` for Windows (using x64 Native Tools Command Prompt for VS, and assuming a gfx1100-compatible AMD GPU): - ```bash - set PATH=%HIP_PATH%\bin;%PATH% - cmake -S . -B build -G Ninja -DAMDGPU_TARGETS=gfx1100 -DGGML_HIPBLAS=ON -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_BUILD_TYPE=Release - cmake --build build - ``` - Make sure that `AMDGPU_TARGETS` is set to the GPU arch you want to compile for. The above example uses `gfx1100` that corresponds to Radeon RX 7900XTX/XT/GRE. You can find a list of targets [here](https://llvm.org/docs/AMDGPUUsage.html#processors) - Find your gpu version string by matching the most significant version information from `rocminfo | grep gfx | head -1 | awk '{print $2}'` with the list of processors, e.g. `gfx1035` maps to `gfx1030`. - - - The environment variable [`HIP_VISIBLE_DEVICES`](https://rocm.docs.amd.com/en/latest/understand/gpu_isolation.html#hip-visible-devices) can be used to specify which GPU(s) will be used. - If your GPU is not officially supported you can use the environment variable [`HSA_OVERRIDE_GFX_VERSION`] set to a similar GPU, for example 10.3.0 on RDNA2 (e.g. gfx1030, gfx1031, or gfx1035) or 11.0.0 on RDNA3. - The following compilation options are also available to tweak performance (yes, they refer to CUDA, not HIP, because it uses the same code as the cuBLAS version above): - - | Option | Legal values | Default | Description | - |------------------------|------------------------|---------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| - | GGML_CUDA_DMMV_X | Positive integer >= 32 | 32 | Number of values in x direction processed by the HIP dequantization + matrix vector multiplication kernel per iteration. Increasing this value can improve performance on fast GPUs. Power of 2 heavily recommended. Does not affect k-quants. | - | GGML_CUDA_MMV_Y | Positive integer | 1 | Block size in y direction for the HIP mul mat vec kernels. Increasing this value can improve performance on fast GPUs. Power of 2 recommended. Does not affect k-quants. | - | GGML_CUDA_KQUANTS_ITER | 1 or 2 | 2 | Number of values processed per iteration and per HIP thread for Q2_K and Q6_K quantization formats. Setting this value to 1 can improve performance for slow GPUs. | - -- #### Vulkan - - **With docker**: - - You don't need to install Vulkan SDK. It will be installed inside the container. - - ```sh - # Build the image - docker build -t llama-cpp-vulkan -f .devops/llama-cli-vulkan.Dockerfile . - - # Then, use it: - docker run -it --rm -v "$(pwd):/app:Z" --device /dev/dri/renderD128:/dev/dri/renderD128 --device /dev/dri/card1:/dev/dri/card1 llama-cpp-vulkan -m "/app/models/YOUR_MODEL_FILE" -p "Building a website can be done in 10 simple steps:" -n 400 -e -ngl 33 - ``` - - **Without docker**: - - Firstly, you need to make sure you have installed [Vulkan SDK](https://vulkan.lunarg.com/doc/view/latest/linux/getting_started_ubuntu.html) - - For example, on Ubuntu 22.04 (jammy), use the command below: - - ```bash - wget -qO - https://packages.lunarg.com/lunarg-signing-key-pub.asc | apt-key add - - wget -qO /etc/apt/sources.list.d/lunarg-vulkan-jammy.list https://packages.lunarg.com/vulkan/lunarg-vulkan-jammy.list - apt update -y - apt-get install -y vulkan-sdk - # To verify the installation, use the command below: - vulkaninfo - ``` - - Alternatively your package manager might be able to provide the appropriate libraries. - For example for Ubuntu 22.04 you can install `libvulkan-dev` instead. - For Fedora 40, you can install `vulkan-devel`, `glslc` and `glslang` packages. - - Then, build llama.cpp using the cmake command below: - - ```bash - cmake -B build -DGGML_VULKAN=1 - cmake --build build --config Release - # Test the output binary (with "-ngl 33" to offload all layers to GPU) - ./bin/llama-cli -m "PATH_TO_MODEL" -p "Hi you how are you" -n 50 -e -ngl 33 -t 4 - - # You should see in the output, ggml_vulkan detected your GPU. For example: - # ggml_vulkan: Using Intel(R) Graphics (ADL GT2) | uma: 1 | fp16: 1 | warp size: 32 - ``` - -### Prepare and Quantize - -> [!NOTE] -> You can use the [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space on Hugging Face to quantise your model weights without any setup too. It is synced from `llama.cpp` main every 6 hours. - -To obtain the official LLaMA 2 weights please see the Obtaining and using the Facebook LLaMA 2 model section. There is also a large selection of pre-quantized `gguf` models available on Hugging Face. - -Note: `convert.py` has been moved to `examples/convert-legacy-llama.py` and shouldn't be used for anything other than `Llama/Llama2/Mistral` models and their derivatives. -It does not support LLaMA 3, you can use `convert-hf-to-gguf.py` with LLaMA 3 downloaded from Hugging Face. +If you want a more ChatGPT-like experience, you can run in conversation mode by passing `-cnv` as a parameter: ```bash -# obtain the official LLaMA model weights and place them in ./models -ls ./models -llama-2-7b tokenizer_checklist.chk tokenizer.model -# [Optional] for models using BPE tokenizers -ls ./models - vocab.json -# [Optional] for PyTorch .bin models like Mistral-7B -ls ./models - +llama-cli -m your_model.gguf -p "You are a helpful assistant" -cnv -# install Python dependencies -python3 -m pip install -r requirements.txt - -# convert the model to ggml FP16 format -python3 convert-hf-to-gguf.py models/mymodel/ - -# quantize the model to 4-bits (using Q4_K_M method) -./llama-quantize ./models/mymodel/ggml-model-f16.gguf ./models/mymodel/ggml-model-Q4_K_M.gguf Q4_K_M - -# update the gguf filetype to current version if older version is now unsupported -./llama-quantize ./models/mymodel/ggml-model-Q4_K_M.gguf ./models/mymodel/ggml-model-Q4_K_M-v2.gguf COPY +# Output: +# > hi, who are you? +# Hi there! I'm your helpful assistant! I'm an AI-powered chatbot designed to assist and provide information to users like you. I'm here to help answer your questions, provide guidance, and offer support on a wide range of topics. I'm a friendly and knowledgeable AI, and I'm always happy to help with anything you need. What's on your mind, and how can I assist you today? +# +# > what is 1+1? +# Easy peasy! The answer to 1+1 is... 2! ``` -### Run the quantized model +By default, the chat template will be taken from the input model. If you want to use another chat template, pass `--chat-template NAME` as a parameter. See the list of [supported templates](https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template) ```bash -# start inference on a gguf model -./llama-cli -m ./models/mymodel/ggml-model-Q4_K_M.gguf -n 128 +./llama-cli -m your_model.gguf -p "You are a helpful assistant" -cnv --chat-template chatml ``` -When running the larger models, make sure you have enough disk space to store all the intermediate files. +You can also use your own template via in-prefix, in-suffix and reverse-prompt parameters: -### Running on Windows with prebuilt binaries - -You will find prebuilt Windows binaries on the release page. - -Simply download and extract the latest zip package of choice: (e.g. `llama-b1380-bin-win-avx2-x64.zip`) - -From the unzipped folder, open a terminal/cmd window here and place a pre-converted `.gguf` model file. Test out the main example like so: - -``` -.\main -m llama-2-7b.Q4_0.gguf -n 128 +```bash +./llama-cli -m your_model.gguf -p "You are a helpful assistant" -cnv --in-prefix 'User: ' --reverse-prompt 'User:' ``` -### Memory/Disk Requirements +### Web server -As the models are currently fully loaded into memory, you will need adequate disk space to save them and sufficient RAM to load them. At the moment, memory and disk requirements are the same. +[llama.cpp web server](./examples/server/README.md) is a lightweight [OpenAI API](https://github.com/openai/openai-openapi) compatible HTTP server that can be used to serve local models and easily connect them to existing clients. -| Model | Original size | Quantized size (Q4_0) | -|------:|--------------:|----------------------:| -| 7B | 13 GB | 3.9 GB | -| 13B | 24 GB | 7.8 GB | -| 30B | 60 GB | 19.5 GB | -| 65B | 120 GB | 38.5 GB | +Example usage: -### Quantization +```bash +./llama-server -m your_model.gguf --port 8080 -Several quantization methods are supported. They differ in the resulting model disk size and inference speed. - -*(outdated)* - -| Model | Measure | F16 | Q4_0 | Q4_1 | Q5_0 | Q5_1 | Q8_0 | -|------:|--------------|-------:|-------:|-------:|-------:|-------:|-------:| -| 7B | perplexity | 5.9066 | 6.1565 | 6.0912 | 5.9862 | 5.9481 | 5.9070 | -| 7B | file size | 13.0G | 3.5G | 3.9G | 4.3G | 4.7G | 6.7G | -| 7B | ms/tok @ 4th | 127 | 55 | 54 | 76 | 83 | 72 | -| 7B | ms/tok @ 8th | 122 | 43 | 45 | 52 | 56 | 67 | -| 7B | bits/weight | 16.0 | 4.5 | 5.0 | 5.5 | 6.0 | 8.5 | -| 13B | perplexity | 5.2543 | 5.3860 | 5.3608 | 5.2856 | 5.2706 | 5.2548 | -| 13B | file size | 25.0G | 6.8G | 7.6G | 8.3G | 9.1G | 13G | -| 13B | ms/tok @ 4th | - | 103 | 105 | 148 | 160 | 131 | -| 13B | ms/tok @ 8th | - | 73 | 82 | 98 | 105 | 128 | -| 13B | bits/weight | 16.0 | 4.5 | 5.0 | 5.5 | 6.0 | 8.5 | - -- [k-quants](https://github.com/ggerganov/llama.cpp/pull/1684) -- recent k-quants improvements and new i-quants - - [#2707](https://github.com/ggerganov/llama.cpp/pull/2707) - - [#2807](https://github.com/ggerganov/llama.cpp/pull/2807) - - [#4773 - 2-bit i-quants (inference)](https://github.com/ggerganov/llama.cpp/pull/4773) - - [#4856 - 2-bit i-quants (inference)](https://github.com/ggerganov/llama.cpp/pull/4856) - - [#4861 - importance matrix](https://github.com/ggerganov/llama.cpp/pull/4861) - - [#4872 - MoE models](https://github.com/ggerganov/llama.cpp/pull/4872) - - [#4897 - 2-bit quantization](https://github.com/ggerganov/llama.cpp/pull/4897) - - [#4930 - imatrix for all k-quants](https://github.com/ggerganov/llama.cpp/pull/4930) - - [#4951 - imatrix on the GPU](https://github.com/ggerganov/llama.cpp/pull/4957) - - [#4969 - imatrix for legacy quants](https://github.com/ggerganov/llama.cpp/pull/4969) - - [#4996 - k-qunats tuning](https://github.com/ggerganov/llama.cpp/pull/4996) - - [#5060 - Q3_K_XS](https://github.com/ggerganov/llama.cpp/pull/5060) - - [#5196 - 3-bit i-quants](https://github.com/ggerganov/llama.cpp/pull/5196) - - [quantization tuning](https://github.com/ggerganov/llama.cpp/pull/5320), [another one](https://github.com/ggerganov/llama.cpp/pull/5334), and [another one](https://github.com/ggerganov/llama.cpp/pull/5361) - -### Perplexity (measuring model quality) - -You can use the `perplexity` example to measure perplexity over a given prompt (lower perplexity is better). -For more information, see [https://huggingface.co/docs/transformers/perplexity](https://huggingface.co/docs/transformers/perplexity). - -The perplexity measurements in table above are done against the `wikitext2` test dataset (https://paperswithcode.com/dataset/wikitext-2), with context length of 512. -The time per token is measured on a MacBook M1 Pro 32GB RAM using 4 and 8 threads. - -#### How to run - -1. Download/extract: https://huggingface.co/datasets/ggml-org/ci/resolve/main/wikitext-2-raw-v1.zip -2. Run `./llama-perplexity -m models/7B/ggml-model-q4_0.gguf -f wiki.test.raw` -3. Output: +# Basic web UI can be accessed via browser: http://localhost:8080 +# Chat completion endpoint: http://localhost:8080/v1/chat/completions ``` -perplexity : calculating perplexity over 655 chunks -24.43 seconds per pass - ETA 4.45 hours -[1]4.5970,[2]5.1807,[3]6.0382,... -``` -And after 4.45 hours, you will have the final perplexity. ### Interactive mode -If you want a more ChatGPT-like experience, you can run in interactive mode by passing `-i` as a parameter. +> [!NOTE] +> If you prefer basic usage, please consider using conversation mode instead of interactive mode + In this mode, you can always interrupt generation by pressing Ctrl+C and entering one or more lines of text, which will be converted into tokens and appended to the current context. You can also specify a *reverse prompt* with the parameter `-r "reverse prompt string"`. This will result in user input being prompted whenever the exact tokens of the reverse prompt string are encountered in the generation. A typical use is to use a prompt that makes LLaMA emulate a chat between multiple users, say Alice and Bob, and pass `-r "Alice:"`. Here is an example of a few-shot interaction, invoked with the command @@ -798,18 +397,71 @@ The `grammars/` folder contains a handful of sample grammars. To write your own, For authoring more complex JSON grammars, you can also check out https://grammar.intrinsiclabs.ai/, a browser app that lets you write TypeScript interfaces which it compiles to GBNF grammars that you can save for local use. Note that the app is built and maintained by members of the community, please file any issues or FRs on [its repo](http://github.com/intrinsiclabsai/gbnfgen) and not this one. -### Obtaining and using the Facebook LLaMA 2 model +## Build -- Refer to [Facebook's LLaMA download page](https://ai.meta.com/resources/models-and-libraries/llama-downloads/) if you want to access the model data. -- Alternatively, if you want to save time and space, you can download already converted and quantized models from [TheBloke](https://huggingface.co/TheBloke), including: - - [LLaMA 2 7B base](https://huggingface.co/TheBloke/Llama-2-7B-GGUF) - - [LLaMA 2 13B base](https://huggingface.co/TheBloke/Llama-2-13B-GGUF) - - [LLaMA 2 70B base](https://huggingface.co/TheBloke/Llama-2-70B-GGUF) - - [LLaMA 2 7B chat](https://huggingface.co/TheBloke/Llama-2-7B-chat-GGUF) - - [LLaMA 2 13B chat](https://huggingface.co/TheBloke/Llama-2-13B-chat-GGUF) - - [LLaMA 2 70B chat](https://huggingface.co/TheBloke/Llama-2-70B-chat-GGUF) +Please refer to [Build llama.cpp locally](./docs/build.md) -### Seminal papers and background on the models +## Supported backends + +| Backend | Target devices | +| --- | --- | +| [Metal](./docs/build.md#metal-build) | Apple Silicon | +| [BLAS](./docs/build.md#blas-build) | All | +| [BLIS](./docs/backend/BLIS.md) | All | +| [SYCL](./docs/backend/SYCL.md) | Intel and Nvidia GPU | +| [MUSA](./docs/build.md#musa) | Moore Threads GPU | +| [CUDA](./docs/build.md#cuda) | Nvidia GPU | +| [hipBLAS](./docs/build.md#hipblas) | AMD GPU | +| [Vulkan](./docs/build.md#vulkan) | GPU | + +## Tools + +### Prepare and Quantize + +> [!NOTE] +> You can use the [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space on Hugging Face to quantise your model weights without any setup too. It is synced from `llama.cpp` main every 6 hours. + +To obtain the official LLaMA 2 weights please see the Obtaining and using the Facebook LLaMA 2 model section. There is also a large selection of pre-quantized `gguf` models available on Hugging Face. + +Note: `convert.py` has been moved to `examples/convert_legacy_llama.py` and shouldn't be used for anything other than `Llama/Llama2/Mistral` models and their derivatives. +It does not support LLaMA 3, you can use `convert_hf_to_gguf.py` with LLaMA 3 downloaded from Hugging Face. + +To learn more about quantizing model, [read this documentation](./examples/quantize/README.md) + +### Perplexity (measuring model quality) + +You can use the `perplexity` example to measure perplexity over a given prompt (lower perplexity is better). +For more information, see [https://huggingface.co/docs/transformers/perplexity](https://huggingface.co/docs/transformers/perplexity). + +To learn more how to measure perplexity using llama.cpp, [read this documentation](./examples/perplexity/README.md) + +## Contributing + +- Contributors can open PRs +- Collaborators can push to branches in the `llama.cpp` repo and merge PRs into the `master` branch +- Collaborators will be invited based on contributions +- Any help with managing issues and PRs is very appreciated! +- See [good first issues](https://github.com/ggerganov/llama.cpp/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) for tasks suitable for first contributions +- Read the [CONTRIBUTING.md](CONTRIBUTING.md) for more information +- Make sure to read this: [Inference at the edge](https://github.com/ggerganov/llama.cpp/discussions/205) +- A bit of backstory for those who are interested: [Changelog podcast](https://changelog.com/podcast/532) + +## Other documentations + +- [main (cli)](./examples/main/README.md) +- [server](./examples/server/README.md) +- [jeopardy](./examples/jeopardy/README.md) +- [GBNF grammars](./grammars/README.md) + +**Development documentations** + +- [How to build](./docs/build.md) +- [Running on Docker](./docs/docker.md) +- [Build on Android](./docs/android.md) +- [Performance troubleshooting](./docs/development/token_generation_performance_tips.md) +- [GGML tips & tricks](https://github.com/ggerganov/llama.cpp/wiki/GGML-Tips-&-Tricks) + +**Seminal papers and background on the models** If your issue is with model generation quality, then please at least scan the following links and papers to understand the limitations of LLaMA models. This is especially important when choosing an appropriate model size and appreciating both the significant and subtle differences between LLaMA models and ChatGPT: - LLaMA: @@ -820,178 +472,3 @@ If your issue is with model generation quality, then please at least scan the fo - GPT-3.5 / InstructGPT / ChatGPT: - [Aligning language models to follow instructions](https://openai.com/research/instruction-following) - [Training language models to follow instructions with human feedback](https://arxiv.org/abs/2203.02155) - -### Android - -#### Build on Android using Termux -[Termux](https://github.com/termux/termux-app#installation) is a method to execute `llama.cpp` on an Android device (no root required). -``` -apt update && apt upgrade -y -apt install git make cmake -``` - -It's recommended to move your model inside the `~/` directory for best performance: -``` -cd storage/downloads -mv model.gguf ~/ -``` - -[Get the code](https://github.com/ggerganov/llama.cpp#get-the-code) & [follow the Linux build instructions](https://github.com/ggerganov/llama.cpp#build) to build `llama.cpp`. - -#### Building the Project using Android NDK -Obtain the [Android NDK](https://developer.android.com/ndk) and then build with CMake. - -Execute the following commands on your computer to avoid downloading the NDK to your mobile. Alternatively, you can also do this in Termux: -``` -$ mkdir build-android -$ cd build-android -$ export NDK= -$ cmake -DCMAKE_TOOLCHAIN_FILE=$NDK/build/cmake/android.toolchain.cmake -DANDROID_ABI=arm64-v8a -DANDROID_PLATFORM=android-23 -DCMAKE_C_FLAGS=-march=armv8.4a+dotprod .. -$ make -``` - -Install [termux](https://github.com/termux/termux-app#installation) on your device and run `termux-setup-storage` to get access to your SD card (if Android 11+ then run the command twice). - -Finally, copy these built `llama` binaries and the model file to your device storage. Because the file permissions in the Android sdcard cannot be changed, you can copy the executable files to the `/data/data/com.termux/files/home/bin` path, and then execute the following commands in Termux to add executable permission: - -(Assumed that you have pushed the built executable files to the /sdcard/llama.cpp/bin path using `adb push`) -``` -$cp -r /sdcard/llama.cpp/bin /data/data/com.termux/files/home/ -$cd /data/data/com.termux/files/home/bin -$chmod +x ./* -``` - -Download model [llama-2-7b-chat.Q4_K_M.gguf](https://huggingface.co/TheBloke/Llama-2-7B-Chat-GGUF/blob/main/llama-2-7b-chat.Q4_K_M.gguf), and push it to `/sdcard/llama.cpp/`, then move it to `/data/data/com.termux/files/home/model/` - -``` -$mv /sdcard/llama.cpp/llama-2-7b-chat.Q4_K_M.gguf /data/data/com.termux/files/home/model/ -``` - -Now, you can start chatting: -``` -$cd /data/data/com.termux/files/home/bin -$./llama-cli -m ../model/llama-2-7b-chat.Q4_K_M.gguf -n 128 -cml -``` - -Here's a demo of an interactive session running on Pixel 5 phone: - -https://user-images.githubusercontent.com/271616/225014776-1d567049-ad71-4ef2-b050-55b0b3b9274c.mp4 - -### Docker - -#### Prerequisites -* Docker must be installed and running on your system. -* Create a folder to store big models & intermediate files (ex. /llama/models) - -#### Images -We have three Docker images available for this project: - -1. `ghcr.io/ggerganov/llama.cpp:full`: This image includes both the main executable file and the tools to convert LLaMA models into ggml and convert into 4-bit quantization. (platforms: `linux/amd64`, `linux/arm64`) -2. `ghcr.io/ggerganov/llama.cpp:light`: This image only includes the main executable file. (platforms: `linux/amd64`, `linux/arm64`) -3. `ghcr.io/ggerganov/llama.cpp:server`: This image only includes the server executable file. (platforms: `linux/amd64`, `linux/arm64`) - -Additionally, there the following images, similar to the above: - -- `ghcr.io/ggerganov/llama.cpp:full-cuda`: Same as `full` but compiled with CUDA support. (platforms: `linux/amd64`) -- `ghcr.io/ggerganov/llama.cpp:light-cuda`: Same as `light` but compiled with CUDA support. (platforms: `linux/amd64`) -- `ghcr.io/ggerganov/llama.cpp:server-cuda`: Same as `server` but compiled with CUDA support. (platforms: `linux/amd64`) -- `ghcr.io/ggerganov/llama.cpp:full-rocm`: Same as `full` but compiled with ROCm support. (platforms: `linux/amd64`, `linux/arm64`) -- `ghcr.io/ggerganov/llama.cpp:light-rocm`: Same as `light` but compiled with ROCm support. (platforms: `linux/amd64`, `linux/arm64`) -- `ghcr.io/ggerganov/llama.cpp:server-rocm`: Same as `server` but compiled with ROCm support. (platforms: `linux/amd64`, `linux/arm64`) - -The GPU enabled images are not currently tested by CI beyond being built. They are not built with any variation from the ones in the Dockerfiles defined in [.devops/](.devops/) and the GitHub Action defined in [.github/workflows/docker.yml](.github/workflows/docker.yml). If you need different settings (for example, a different CUDA or ROCm library, you'll need to build the images locally for now). - -#### Usage - -The easiest way to download the models, convert them to ggml and optimize them is with the --all-in-one command which includes the full docker image. - -Replace `/path/to/models` below with the actual path where you downloaded the models. - -```bash -docker run -v /path/to/models:/models ghcr.io/ggerganov/llama.cpp:full --all-in-one "/models/" 7B -``` - -On completion, you are ready to play! - -```bash -docker run -v /path/to/models:/models ghcr.io/ggerganov/llama.cpp:full --run -m /models/7B/ggml-model-q4_0.gguf -p "Building a website can be done in 10 simple steps:" -n 512 -``` - -or with a light image: - -```bash -docker run -v /path/to/models:/models ghcr.io/ggerganov/llama.cpp:light -m /models/7B/ggml-model-q4_0.gguf -p "Building a website can be done in 10 simple steps:" -n 512 -``` - -or with a server image: - -```bash -docker run -v /path/to/models:/models -p 8000:8000 ghcr.io/ggerganov/llama.cpp:server -m /models/7B/ggml-model-q4_0.gguf --port 8000 --host 0.0.0.0 -n 512 -``` - -### Docker With CUDA - -Assuming one has the [nvidia-container-toolkit](https://github.com/NVIDIA/nvidia-container-toolkit) properly installed on Linux, or is using a GPU enabled cloud, `cuBLAS` should be accessible inside the container. - -#### Building Locally - -```bash -docker build -t local/llama.cpp:full-cuda -f .devops/full-cuda.Dockerfile . -docker build -t local/llama.cpp:light-cuda -f .devops/llama-cli-cuda.Dockerfile . -docker build -t local/llama.cpp:server-cuda -f .devops/llama-server-cuda.Dockerfile . -``` - -You may want to pass in some different `ARGS`, depending on the CUDA environment supported by your container host, as well as the GPU architecture. - -The defaults are: - -- `CUDA_VERSION` set to `11.7.1` -- `CUDA_DOCKER_ARCH` set to `all` - -The resulting images, are essentially the same as the non-CUDA images: - -1. `local/llama.cpp:full-cuda`: This image includes both the main executable file and the tools to convert LLaMA models into ggml and convert into 4-bit quantization. -2. `local/llama.cpp:light-cuda`: This image only includes the main executable file. -3. `local/llama.cpp:server-cuda`: This image only includes the server executable file. - -#### Usage - -After building locally, Usage is similar to the non-CUDA examples, but you'll need to add the `--gpus` flag. You will also want to use the `--n-gpu-layers` flag. - -```bash -docker run --gpus all -v /path/to/models:/models local/llama.cpp:full-cuda --run -m /models/7B/ggml-model-q4_0.gguf -p "Building a website can be done in 10 simple steps:" -n 512 --n-gpu-layers 1 -docker run --gpus all -v /path/to/models:/models local/llama.cpp:light-cuda -m /models/7B/ggml-model-q4_0.gguf -p "Building a website can be done in 10 simple steps:" -n 512 --n-gpu-layers 1 -docker run --gpus all -v /path/to/models:/models local/llama.cpp:server-cuda -m /models/7B/ggml-model-q4_0.gguf --port 8000 --host 0.0.0.0 -n 512 --n-gpu-layers 1 -``` - -### Contributing - -- Contributors can open PRs -- Collaborators can push to branches in the `llama.cpp` repo and merge PRs into the `master` branch -- Collaborators will be invited based on contributions -- Any help with managing issues and PRs is very appreciated! -- Make sure to read this: [Inference at the edge](https://github.com/ggerganov/llama.cpp/discussions/205) -- A bit of backstory for those who are interested: [Changelog podcast](https://changelog.com/podcast/532) - -### Coding guidelines - -- Avoid adding third-party dependencies, extra files, extra headers, etc. -- Always consider cross-compatibility with other operating systems and architectures -- Avoid fancy looking modern STL constructs, use basic `for` loops, avoid templates, keep it simple -- There are no strict rules for the code style, but try to follow the patterns in the code (indentation, spaces, etc.). Vertical alignment makes things more readable and easier to batch edit -- Clean-up any trailing whitespaces, use 4 spaces for indentation, brackets on the same line, `void * ptr`, `int & a` -- See [good first issues](https://github.com/ggerganov/llama.cpp/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) for tasks suitable for first contributions -- Tensors store data in row-major order. We refer to dimension 0 as columns, 1 as rows, 2 as matrices -- Matrix multiplication is unconventional: [`C = ggml_mul_mat(ctx, A, B)`](https://github.com/ggerganov/llama.cpp/blob/880e352277fc017df4d5794f0c21c44e1eae2b84/ggml.h#L1058-L1064) means $C^T = A B^T \Leftrightarrow C = B A^T.$ - -![matmul](media/matmul.png) - -### Docs - -- [main (cli)](./examples/main/README.md) -- [server](./examples/server/README.md) -- [jeopardy](./examples/jeopardy/README.md) -- [BLIS](./docs/BLIS.md) -- [Performance troubleshooting](./docs/token_generation_performance_tips.md) -- [GGML tips & tricks](https://github.com/ggerganov/llama.cpp/wiki/GGML-Tips-&-Tricks) -- [GBNF grammars](./grammars/README.md) diff --git a/ci/run.sh b/ci/run.sh index e0cedb24f..58022c7dc 100755 --- a/ci/run.sh +++ b/ci/run.sh @@ -103,6 +103,9 @@ function gg_run_ctest_debug { set -e + # Check cmake, make and ctest are installed + gg_check_build_requirements + (time cmake -DCMAKE_BUILD_TYPE=Debug ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log (time make -j ) 2>&1 | tee -a $OUT/${ci}-make.log @@ -131,6 +134,9 @@ function gg_run_ctest_release { set -e + # Check cmake, make and ctest are installed + gg_check_build_requirements + (time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log (time make -j ) 2>&1 | tee -a $OUT/${ci}-make.log @@ -287,7 +293,7 @@ function gg_run_open_llama_7b_v2 { (time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} -DGGML_CUDA=1 .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log (time make -j ) 2>&1 | tee -a $OUT/${ci}-make.log - python3 ../examples/convert-legacy-llama.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf + python3 ../examples/convert_legacy_llama.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf model_f16="${path_models}/ggml-model-f16.gguf" model_q8_0="${path_models}/ggml-model-q8_0.gguf" @@ -421,7 +427,7 @@ function gg_run_pythia_1_4b { (time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log (time make -j ) 2>&1 | tee -a $OUT/${ci}-make.log - python3 ../convert-hf-to-gguf.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf + python3 ../convert_hf_to_gguf.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf model_f16="${path_models}/ggml-model-f16.gguf" model_q8_0="${path_models}/ggml-model-q8_0.gguf" @@ -553,7 +559,7 @@ function gg_run_pythia_2_8b { (time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} -DGGML_CUDA=1 .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log (time make -j ) 2>&1 | tee -a $OUT/${ci}-make.log - python3 ../convert-hf-to-gguf.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf + python3 ../convert_hf_to_gguf.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf model_f16="${path_models}/ggml-model-f16.gguf" model_q8_0="${path_models}/ggml-model-q8_0.gguf" @@ -688,7 +694,7 @@ function gg_run_embd_bge_small { (time cmake -DCMAKE_BUILD_TYPE=Release ${CMAKE_EXTRA} .. ) 2>&1 | tee -a $OUT/${ci}-cmake.log (time make -j ) 2>&1 | tee -a $OUT/${ci}-make.log - python3 ../convert-hf-to-gguf.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf + python3 ../convert_hf_to_gguf.py ${path_models} --outfile ${path_models}/ggml-model-f16.gguf model_f16="${path_models}/ggml-model-f16.gguf" model_q8_0="${path_models}/ggml-model-q8_0.gguf" @@ -701,6 +707,20 @@ function gg_run_embd_bge_small { set +e } +function gg_check_build_requirements { + if ! command -v cmake &> /dev/null; then + gg_printf 'cmake not found, please install' + fi + + if ! command -v make &> /dev/null; then + gg_printf 'make not found, please install' + fi + + if ! command -v ctest &> /dev/null; then + gg_printf 'ctest not found, please install' + fi +} + function gg_sum_embd_bge_small { gg_printf '### %s\n\n' "${ci}" diff --git a/cmake/llama-config.cmake.in b/cmake/llama-config.cmake.in index 2e7da2f8e..f072b76a3 100644 --- a/cmake/llama-config.cmake.in +++ b/cmake/llama-config.cmake.in @@ -8,6 +8,13 @@ set(GGML_CUDA @GGML_CUDA@) set(GGML_METAL @GGML_METAL@) set(GGML_HIPBLAS @GGML_HIPBLAS@) set(GGML_ACCELERATE @GGML_ACCELERATE@) +set(GGML_VULKAN @GGML_VULKAN@) +set(GGML_VULKAN_CHECK_RESULTS @GGML_VULKAN_CHECK_RESULTS@) +set(GGML_VULKAN_DEBUG @GGML_VULKAN_DEBUG@) +set(GGML_VULKAN_MEMORY_DEBUG @GGML_VULKAN_MEMORY_DEBUG@) +set(GGML_VULKAN_VALIDATE @GGML_VULKAN_VALIDATE@) +set(GGML_SYCL @GGML_SYCL@) +set(GGML_OPENMP @GGML_OPENMP@) @PACKAGE_INIT@ @@ -37,18 +44,36 @@ if (GGML_METAL) find_library(METALKIT_FRAMEWORK MetalKit REQUIRED) endif() +if (GGML_VULKAN) + find_package(Vulkan REQUIRED) +endif() + if (GGML_HIPBLAS) find_package(hip REQUIRED) find_package(hipblas REQUIRED) find_package(rocblas REQUIRED) endif() +if (GGML_SYCL) + find_package(IntelSYCL REQUIRED) + find_package(MKL REQUIRED) +endif() + +if (GGML_OPENMP) + find_package(OpenMP REQUIRED) +endif() + + +find_library(ggml_LIBRARY ggml + REQUIRED + HINTS ${LLAMA_LIB_DIR}) + find_library(llama_LIBRARY llama REQUIRED HINTS ${LLAMA_LIB_DIR}) -set(_llama_link_deps "Threads::Threads" "@LLAMA_EXTRA_LIBS@") -set(_llama_transient_defines "@LLAMA_TRANSIENT_DEFINES@") +set(_llama_link_deps "${ggml_LIBRARY}" "@GGML_LINK_LIBRARIES@") +set(_llama_transient_defines "@GGML_TRANSIENT_DEFINES@") add_library(llama UNKNOWN IMPORTED) diff --git a/common/common.cpp b/common/common.cpp index c76d0e2c3..60c7eac75 100644 --- a/common/common.cpp +++ b/common/common.cpp @@ -1,3 +1,7 @@ +#if defined(_MSC_VER) +#define _SILENCE_CXX17_CODECVT_HEADER_DEPRECATION_WARNING +#endif + #include "common.h" // Change JSON_ASSERT from assert() to GGML_ASSERT: #define JSON_ASSERT GGML_ASSERT @@ -190,6 +194,12 @@ int32_t cpu_get_num_math() { // CLI argument parsing // +void gpt_params_handle_hf_token(gpt_params & params) { + if (params.hf_token.empty() && std::getenv("HF_TOKEN")) { + params.hf_token = std::getenv("HF_TOKEN"); + } +} + void gpt_params_handle_model_default(gpt_params & params) { if (!params.hf_repo.empty()) { // short-hand to avoid specifying --hf-file -> default it to --model @@ -237,6 +247,8 @@ bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) { gpt_params_handle_model_default(params); + gpt_params_handle_hf_token(params); + if (params.escape) { string_process_escapes(params.prompt); string_process_escapes(params.input_prefix); @@ -472,6 +484,14 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa else { invalid_param = true; } return true; } + if (arg == "--attention") { + CHECK_ARG + std::string value(argv[i]); + /**/ if (value == "causal") { params.attention_type = LLAMA_ATTENTION_TYPE_CAUSAL; } + else if (value == "non-causal") { params.attention_type = LLAMA_ATTENTION_TYPE_NON_CAUSAL; } + else { invalid_param = true; } + return true; + } if (arg == "--defrag-thold" || arg == "-dt") { CHECK_ARG params.defrag_thold = std::stof(argv[i]); @@ -644,6 +664,14 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa params.model_url = argv[i]; return true; } + if (arg == "-hft" || arg == "--hf-token") { + if (++i >= argc) { + invalid_param = true; + return true; + } + params.hf_token = argv[i]; + return true; + } if (arg == "-hfr" || arg == "--hf-repo") { CHECK_ARG params.hf_repo = argv[i]; @@ -657,7 +685,6 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa if (arg == "--lora") { CHECK_ARG params.lora_adapter.emplace_back(argv[i], 1.0f); - params.use_mmap = false; return true; } if (arg == "--lora-scaled") { @@ -665,12 +692,6 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa const char* lora_adapter = argv[i]; CHECK_ARG params.lora_adapter.emplace_back(lora_adapter, std::stof(argv[i])); - params.use_mmap = false; - return true; - } - if (arg == "--lora-base") { - CHECK_ARG - params.lora_base = argv[i]; return true; } if (arg == "--control-vector") { @@ -757,7 +778,7 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa params.cache_type_v = argv[++i]; return true; } - if (arg == "--multiline-input") { + if (arg == "-mli" || arg == "--multiline-input") { params.multiline_input = true; return true; } @@ -769,6 +790,10 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa params.cont_batching = true; return true; } + if (arg == "-nocb" || arg == "--no-cont-batching") { + params.cont_batching = false; + return true; + } if (arg == "-fa" || arg == "--flash-attn") { params.flash_attn = true; return true; @@ -1014,16 +1039,23 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa } if (arg == "--in-prefix-bos") { params.input_prefix_bos = true; + params.enable_chat_template = false; return true; } if (arg == "--in-prefix") { CHECK_ARG params.input_prefix = argv[i]; + params.enable_chat_template = false; return true; } if (arg == "--in-suffix") { CHECK_ARG params.input_suffix = argv[i]; + params.enable_chat_template = false; + return true; + } + if (arg == "--spm-infill") { + params.spm_infill = true; return true; } if (arg == "--grammar") { @@ -1237,6 +1269,7 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa CHECK_ARG params.out_file = argv[i]; params.cvector_outfile = argv[i]; + params.lora_outfile = argv[i]; return true; } if (arg == "-ofreq" || arg == "--output-frequency") { @@ -1291,6 +1324,10 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa else { invalid_param = true; } return true; } + if (arg == "--no-warmup") { + params.warmup = false; + return true; + } #ifndef LOG_DISABLE_LOGS // Parse args for logging parameters if (log_param_single_parse(argv[i])) { @@ -1387,7 +1424,9 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param options.push_back({ "*", " --keep N", "number of tokens to keep from the initial prompt (default: %d, -1 = all)", params.n_keep }); options.push_back({ "*", " --chunks N", "max number of chunks to process (default: %d, -1 = all)", params.n_chunks }); options.push_back({ "*", "-fa, --flash-attn", "enable Flash Attention (default: %s)", params.flash_attn ? "enabled" : "disabled" }); - options.push_back({ "*", "-p, --prompt PROMPT", "prompt to start generation with (default: '%s')", params.prompt.c_str() }); + options.push_back({ "*", "-p, --prompt PROMPT", "prompt to start generation with\n" + "in conversation mode, this will be used as system prompt\n" + "(default: '%s')", params.prompt.c_str() }); options.push_back({ "*", "-f, --file FNAME", "a file containing the prompt (default: none)" }); options.push_back({ "*", " --in-file FNAME", "an input file (repeat to specify multiple files)" }); options.push_back({ "*", "-bf, --binary-file FNAME", "binary file containing the prompt (default: none)" }); @@ -1402,13 +1441,18 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param "halt generation at PROMPT, return control in interactive mode\n" "can be specified more than once for multiple prompts" }); options.push_back({ "main", "-sp, --special", "special tokens output enabled (default: %s)", params.special ? "true" : "false" }); - options.push_back({ "main", "-cnv, --conversation", "run in conversation mode (does not print special tokens and suffix/prefix) (default: %s)", params.conversation ? "true" : "false" }); + options.push_back({ "main", "-cnv, --conversation", "run in conversation mode, does not print special tokens and suffix/prefix\n" + "if suffix/prefix are not specified, default chat template will be used\n" + "(default: %s)", params.conversation ? "true" : "false" }); options.push_back({ "main infill", "-i, --interactive", "run in interactive mode (default: %s)", params.interactive ? "true" : "false" }); options.push_back({ "main infill", "-if, --interactive-first", "run in interactive mode and wait for input right away (default: %s)", params.interactive_first ? "true" : "false" }); options.push_back({ "main infill", "-mli, --multiline-input", "allows you to write or paste multiple lines without ending each in '\\'" }); options.push_back({ "main infill", " --in-prefix-bos", "prefix BOS to user inputs, preceding the `--in-prefix` string" }); options.push_back({ "main infill", " --in-prefix STRING", "string to prefix user inputs with (default: empty)" }); options.push_back({ "main infill", " --in-suffix STRING", "string to suffix after user inputs with (default: empty)" }); + options.push_back({ "main", " --no-warmup", "skip warming up the model with an empty run" }); + options.push_back({ "server infill", + " --spm-infill", "use Suffix/Prefix/Middle pattern for infill (instead of Prefix/Suffix/Middle) as some models prefer this. (default: %s)", params.spm_infill ? "enabled" : "disabled" }); options.push_back({ "sampling" }); options.push_back({ "*", " --samplers SAMPLERS", "samplers that will be used for generation in the order, separated by \';\'\n" @@ -1444,6 +1488,7 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param options.push_back({ "main", " --cfg-scale N", "strength of guidance (default: %.1f, 1.0 = disable)", (double)sparams.cfg_scale }); options.push_back({ "main", " --chat-template JINJA_TEMPLATE", "set custom jinja chat template (default: template taken from model's metadata)\n" + "if suffix/prefix are specified, template will be disabled\n" "only commonly used templates are accepted:\n" "https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template" }); options.push_back({ "grammar" }); @@ -1454,8 +1499,10 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param "For schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead" }); options.push_back({ "embedding" }); - options.push_back({ "embedding", " --pooling {none,mean,cls}", + options.push_back({ "embedding", " --pooling {none,mean,cls,last}", "pooling type for embeddings, use model default if unspecified" }); + options.push_back({ "embedding", " --attention {causal,non-causal}", + "attention type for embeddings, use model default if unspecified" }); options.push_back({ "context hacking" }); options.push_back({ "*", " --rope-scaling {none,linear,yarn}", @@ -1494,6 +1541,7 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param options.push_back({ "*", "-np, --parallel N", "number of parallel sequences to decode (default: %d)", params.n_parallel }); options.push_back({ "*", "-ns, --sequences N", "number of sequences to decode (default: %d)", params.n_sequences }); options.push_back({ "*", "-cb, --cont-batching", "enable continuous batching (a.k.a dynamic batching) (default: %s)", params.cont_batching ? "enabled" : "disabled" }); + options.push_back({ "*", "-nocb, --no-cont-batching", "disable continuous batching" }); options.push_back({ "multi-modality" }); options.push_back({ "*", " --mmproj FILE", "path to a multimodal projector file for LLaVA. see examples/llava/README.md" }); @@ -1536,9 +1584,8 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param options.push_back({ "*", " --override-kv KEY=TYPE:VALUE", "advanced option to override model metadata by key. may be specified multiple times.\n" "types: int, float, bool, str. example: --override-kv tokenizer.ggml.add_bos_token=bool:false" }); - options.push_back({ "*", " --lora FNAME", "apply LoRA adapter (implies --no-mmap)" }); - options.push_back({ "*", " --lora-scaled FNAME S", "apply LoRA adapter with user defined scaling S (implies --no-mmap)" }); - options.push_back({ "*", " --lora-base FNAME", "optional model to use as a base for the layers modified by the LoRA adapter" }); + options.push_back({ "*", " --lora FNAME", "apply LoRA adapter (can be repeated to use multiple adapters)" }); + options.push_back({ "*", " --lora-scaled FNAME S", "apply LoRA adapter with user defined scaling S (can be repeated to use multiple adapters)" }); options.push_back({ "*", " --control-vector FNAME", "add a control vector\n" "note: this argument can be repeated to add multiple control vectors" }); options.push_back({ "*", " --control-vector-scaled FNAME SCALE", @@ -1552,6 +1599,7 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param options.push_back({ "*", "-mu, --model-url MODEL_URL", "model download url (default: unused)" }); options.push_back({ "*", "-hfr, --hf-repo REPO", "Hugging Face model repository (default: unused)" }); options.push_back({ "*", "-hff, --hf-file FILE", "Hugging Face model file (default: unused)" }); + options.push_back({ "*", "-hft, --hf-token TOKEN", "Hugging Face access token (default: value from HF_TOKEN environment variable)" }); options.push_back({ "retrieval" }); options.push_back({ "retrieval", " --context-file FNAME", "file to load context from (repeat to specify multiple files)" }); @@ -1628,6 +1676,13 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param options.push_back({ "cvector", " --pca-iter N", "number of iterations used for PCA (default: %d)", params.n_pca_iterations }); options.push_back({ "cvector", " --method {pca,mean}", "dimensionality reduction method to be used (default: pca)" }); + options.push_back({ "export-lora" }); + options.push_back({ "export-lora", "-m, --model", "model path from which to load base model (default '%s')", params.model.c_str() }); + options.push_back({ "export-lora", " --lora FNAME", "path to LoRA adapter (can be repeated to use multiple adapters)" }); + options.push_back({ "export-lora", " --lora-scaled FNAME S", "path to LoRA adapter with user defined scaling S (can be repeated to use multiple adapters)" }); + options.push_back({ "*", "-t, --threads N", "number of threads to use during computation (default: %d)", params.n_threads }); + options.push_back({ "export-lora", "-o, --output FNAME", "output file (default: '%s')", params.lora_outfile.c_str() }); + printf("usage: %s [options]\n", argv[0]); for (const auto & o : options) { @@ -1991,9 +2046,9 @@ std::tuple llama_init_from_gpt_par llama_model * model = nullptr; if (!params.hf_repo.empty() && !params.hf_file.empty()) { - model = llama_load_model_from_hf(params.hf_repo.c_str(), params.hf_file.c_str(), params.model.c_str(), mparams); + model = llama_load_model_from_hf(params.hf_repo.c_str(), params.hf_file.c_str(), params.model.c_str(), params.hf_token.c_str(), mparams); } else if (!params.model_url.empty()) { - model = llama_load_model_from_url(params.model_url.c_str(), params.model.c_str(), mparams); + model = llama_load_model_from_url(params.model_url.c_str(), params.model.c_str(), params.hf_token.c_str(), mparams); } else { model = llama_load_model_from_file(params.model.c_str(), mparams); } @@ -2039,19 +2094,14 @@ std::tuple llama_init_from_gpt_par for (unsigned int i = 0; i < params.lora_adapter.size(); ++i) { const std::string & lora_adapter = std::get<0>(params.lora_adapter[i]); float lora_scale = std::get<1>(params.lora_adapter[i]); - int err = llama_model_apply_lora_from_file(model, - lora_adapter.c_str(), - lora_scale, - ((i > 0) || params.lora_base.empty()) - ? NULL - : params.lora_base.c_str(), - params.n_threads); - if (err != 0) { + auto adapter = llama_lora_adapter_init(model, lora_adapter.c_str()); + if (adapter == nullptr) { fprintf(stderr, "%s: error: failed to apply lora adapter\n", __func__); llama_free(lctx); llama_free_model(model); return std::make_tuple(nullptr, nullptr); } + llama_lora_adapter_set(lctx, adapter, lora_scale); } if (params.ignore_eos) { @@ -2061,7 +2111,24 @@ std::tuple llama_init_from_gpt_par if (params.warmup) { LOG("warming up the model with an empty run\n"); - std::vector tmp = { llama_token_bos(model), llama_token_eos(model), }; + std::vector tmp; + llama_token bos = llama_token_bos(model); + llama_token eos = llama_token_eos(model); + // some models (e.g. T5) don't have a BOS token + if (bos != -1) { + tmp.push_back(bos); + } + tmp.push_back(eos); + + if (llama_model_has_encoder(model)) { + llama_encode(lctx, llama_batch_get_one(tmp.data(), tmp.size(), 0, 0)); + llama_token decoder_start_token_id = llama_model_decoder_start_token(model); + if (decoder_start_token_id == -1) { + decoder_start_token_id = bos; + } + tmp.clear(); + tmp.push_back(decoder_start_token_id); + } llama_decode(lctx, llama_batch_get_one(tmp.data(), std::min(tmp.size(), (size_t) params.n_batch), 0, 0)); llama_kv_cache_clear(lctx); llama_synchronize(lctx); @@ -2144,6 +2211,7 @@ struct llama_context_params llama_context_params_from_gpt_params(const gpt_param cparams.yarn_beta_slow = params.yarn_beta_slow; cparams.yarn_orig_ctx = params.yarn_orig_ctx; cparams.pooling_type = params.pooling_type; + cparams.attention_type = params.attention_type; cparams.defrag_thold = params.defrag_thold; cparams.cb_eval = params.cb_eval; cparams.cb_eval_user_data = params.cb_eval_user_data; @@ -2163,7 +2231,7 @@ static bool starts_with(const std::string & str, const std::string & prefix) { return str.rfind(prefix, 0) == 0; } -static bool llama_download_file(const std::string & url, const std::string & path) { +static bool llama_download_file(const std::string & url, const std::string & path, const std::string & hf_token) { // Initialize libcurl std::unique_ptr curl(curl_easy_init(), &curl_easy_cleanup); @@ -2178,6 +2246,15 @@ static bool llama_download_file(const std::string & url, const std::string & pat curl_easy_setopt(curl.get(), CURLOPT_URL, url.c_str()); curl_easy_setopt(curl.get(), CURLOPT_FOLLOWLOCATION, 1L); + // Check if hf-token or bearer-token was specified + if (!hf_token.empty()) { + std::string auth_header = "Authorization: Bearer "; + auth_header += hf_token.c_str(); + struct curl_slist *http_headers = NULL; + http_headers = curl_slist_append(http_headers, auth_header.c_str()); + curl_easy_setopt(curl.get(), CURLOPT_HTTPHEADER, http_headers); + } + #if defined(_WIN32) // CURLSSLOPT_NATIVE_CA tells libcurl to use standard certificate store of // operating system. Currently implemented under MS-Windows. @@ -2373,6 +2450,7 @@ static bool llama_download_file(const std::string & url, const std::string & pat struct llama_model * llama_load_model_from_url( const char * model_url, const char * path_model, + const char * hf_token, const struct llama_model_params & params) { // Basic validation of the model_url if (!model_url || strlen(model_url) == 0) { @@ -2380,7 +2458,7 @@ struct llama_model * llama_load_model_from_url( return NULL; } - if (!llama_download_file(model_url, path_model)) { + if (!llama_download_file(model_url, path_model, hf_token)) { return NULL; } @@ -2428,14 +2506,14 @@ struct llama_model * llama_load_model_from_url( // Prepare download in parallel std::vector> futures_download; for (int idx = 1; idx < n_split; idx++) { - futures_download.push_back(std::async(std::launch::async, [&split_prefix, &split_url_prefix, &n_split](int download_idx) -> bool { + futures_download.push_back(std::async(std::launch::async, [&split_prefix, &split_url_prefix, &n_split, hf_token](int download_idx) -> bool { char split_path[PATH_MAX] = {0}; llama_split_path(split_path, sizeof(split_path), split_prefix, download_idx, n_split); char split_url[LLAMA_CURL_MAX_URL_LENGTH] = {0}; llama_split_path(split_url, sizeof(split_url), split_url_prefix, download_idx, n_split); - return llama_download_file(split_url, split_path); + return llama_download_file(split_url, split_path, hf_token); }, idx)); } @@ -2454,6 +2532,7 @@ struct llama_model * llama_load_model_from_hf( const char * repo, const char * model, const char * path_model, + const char * hf_token, const struct llama_model_params & params) { // construct hugging face model url: // @@ -2469,7 +2548,7 @@ struct llama_model * llama_load_model_from_hf( model_url += "/resolve/main/"; model_url += model; - return llama_load_model_from_url(model_url.c_str(), path_model, params); + return llama_load_model_from_url(model_url.c_str(), path_model, hf_token, params); } #else @@ -2477,6 +2556,7 @@ struct llama_model * llama_load_model_from_hf( struct llama_model * llama_load_model_from_url( const char * /*model_url*/, const char * /*path_model*/, + const char * /*hf_token*/, const struct llama_model_params & /*params*/) { fprintf(stderr, "%s: llama.cpp built without libcurl, downloading from an url not supported.\n", __func__); return nullptr; @@ -2486,6 +2566,7 @@ struct llama_model * llama_load_model_from_hf( const char * /*repo*/, const char * /*model*/, const char * /*path_model*/, + const char * /*hf_token*/, const struct llama_model_params & /*params*/) { fprintf(stderr, "%s: llama.cpp built without libcurl, downloading from Hugging Face not supported.\n", __func__); return nullptr; @@ -2550,51 +2631,35 @@ std::vector llama_tokenize( } std::string llama_token_to_piece(const struct llama_context * ctx, llama_token token, bool special) { - std::vector result(8, 0); - const int n_tokens = llama_token_to_piece(llama_get_model(ctx), token, result.data(), result.size(), special); - if (n_tokens < 0) { - result.resize(-n_tokens); - int check = llama_token_to_piece(llama_get_model(ctx), token, result.data(), result.size(), special); - GGML_ASSERT(check == -n_tokens); - } else { - result.resize(n_tokens); + std::string piece; + piece.resize(piece.capacity()); // using string internal cache, 15 bytes + '\n' + const int n_chars = llama_token_to_piece(llama_get_model(ctx), token, &piece[0], piece.size(), 0, special); + if (n_chars < 0) { + piece.resize(-n_chars); + int check = llama_token_to_piece(llama_get_model(ctx), token, &piece[0], piece.size(), 0, special); + GGML_ASSERT(check == -n_chars); + } + else { + piece.resize(n_chars); } - return std::string(result.data(), result.size()); + return piece; } -std::string llama_detokenize_spm(llama_context * ctx, const std::vector & tokens) { - const llama_token bos_id = llama_token_bos(llama_get_model(ctx)); - - std::string piece; - std::string result; - - for (size_t i = 0; i < tokens.size(); ++i) { - piece = llama_token_to_piece(ctx, tokens[i]); - - // remove the leading space of the first non-BOS token - if (((tokens[0] == bos_id && i == 1) || (tokens[0] != bos_id && i == 0)) && piece[0] == ' ') { - piece = piece.substr(1); - } - - result += piece; +std::string llama_detokenize(llama_context * ctx, const std::vector & tokens, bool special) { + std::string text; + text.resize(std::max(text.capacity(), tokens.size())); + int32_t n_chars = llama_detokenize(llama_get_model(ctx), tokens.data(), (int32_t)tokens.size(), &text[0], (int32_t)text.size(), false, special); + if (n_chars < 0) { + text.resize(-n_chars); + n_chars = llama_detokenize(llama_get_model(ctx), tokens.data(), (int32_t)tokens.size(), &text[0], (int32_t)text.size(), false, special); + GGML_ASSERT(n_chars <= (int32_t)text.size()); // whitespace trimming is performed after per-token detokenization } - return result; -} - -std::string llama_detokenize_bpe(llama_context * ctx, const std::vector & tokens) { - std::string piece; - std::string result; - - for (size_t i = 0; i < tokens.size(); ++i) { - piece = llama_token_to_piece(ctx, tokens[i]); - - result += piece; - } + text.resize(n_chars); // NOTE: the original tokenizer decodes bytes after collecting the pieces. - return result; + return text; } bool llama_should_add_bos_token(const llama_model * model) { @@ -2618,6 +2683,7 @@ std::string llama_chat_apply_template(const struct llama_model * model, const std::vector & msgs, bool add_ass) { int alloc_size = 0; + bool fallback = false; // indicate if we must fallback to default chatml std::vector chat; for (auto & msg : msgs) { chat.push_back({msg.role.c_str(), msg.content.c_str()}); @@ -2630,10 +2696,26 @@ std::string llama_chat_apply_template(const struct llama_model * model, // run the first time to get the total output length int32_t res = llama_chat_apply_template(model, ptr_tmpl, chat.data(), chat.size(), add_ass, buf.data(), buf.size()); + // error: chat template is not supported + if (res < 0) { + if (ptr_tmpl != nullptr) { + // if the custom "tmpl" is not supported, we throw an error + // this is a bit redundant (for good), since we're not sure if user validated the custom template with llama_chat_verify_template() + throw std::runtime_error("this custom template is not supported"); + } else { + // If the built-in template is not supported, we default to chatml + res = llama_chat_apply_template(nullptr, "chatml", chat.data(), chat.size(), add_ass, buf.data(), buf.size()); + fallback = true; + } + } + // if it turns out that our buffer is too small, we resize it if ((size_t) res > buf.size()) { buf.resize(res); - res = llama_chat_apply_template(model, ptr_tmpl, chat.data(), chat.size(), add_ass, buf.data(), buf.size()); + res = llama_chat_apply_template( + fallback ? nullptr : model, + fallback ? "chatml" : ptr_tmpl, + chat.data(), chat.size(), add_ass, buf.data(), buf.size()); } std::string formatted_chat(buf.data(), res); @@ -2645,12 +2727,19 @@ std::string llama_chat_format_single(const struct llama_model * model, const std::vector & past_msg, const llama_chat_msg & new_msg, bool add_ass) { - auto fmt_past_msg = llama_chat_apply_template(model, tmpl, past_msg, false); + std::ostringstream ss; + auto fmt_past_msg = past_msg.empty() ? "" : llama_chat_apply_template(model, tmpl, past_msg, false); std::vector chat_new(past_msg); + // if the past_msg ends with a newline, we must preserve it in the formatted version + if (add_ass && !fmt_past_msg.empty() && fmt_past_msg.back() == '\n') { + ss << "\n"; + }; + // format chat with new_msg chat_new.push_back(new_msg); auto fmt_new_msg = llama_chat_apply_template(model, tmpl, chat_new, add_ass); - auto formatted = fmt_new_msg.substr(fmt_past_msg.size(), fmt_new_msg.size() - fmt_past_msg.size()); - return formatted; + // get the diff part + ss << fmt_new_msg.substr(fmt_past_msg.size(), fmt_new_msg.size() - fmt_past_msg.size()); + return ss.str(); } std::string llama_chat_format_example(const struct llama_model * model, @@ -2804,125 +2893,87 @@ float llama_embd_similarity_cos(const float * embd1, const float * embd2, int n) // static llama_control_vector_data llama_control_vector_load_one(const llama_control_vector_load_info & load_info) { - int32_t n_tensors; - - size_t n_bytes = 0; - - uint32_t max_direction_layer = 0; - llama_control_vector_data result = { -1, {} }; - // calculate size of ctx needed for tensors, ensure tensors are f32, and find max layer - { - struct ggml_init_params meta_params = { - /* .mem_size = */ ggml_tensor_overhead() * 128 + ggml_graph_overhead(), - /* .mem_buffer = */ nullptr, - /* .no_alloc = */ true, - }; - ggml_context * meta_ctx = ggml_init(meta_params); - struct gguf_init_params meta_gguf_params = { - /* .no_alloc = */ true, - /* .ctx = */ &meta_ctx, - }; - struct gguf_context * meta_ctx_gguf = gguf_init_from_file(load_info.fname.c_str(), meta_gguf_params); - if (!meta_ctx_gguf) { - fprintf(stderr, "%s: failed to load control vector from %s\n", __func__, load_info.fname.c_str()); - ggml_free(meta_ctx); - return result; - } - - n_tensors = gguf_get_n_tensors(meta_ctx_gguf); - for (int i = 0; i < n_tensors; i++) { - std::string name = gguf_get_tensor_name(meta_ctx_gguf, i); - - // split on '.' - size_t dotpos = name.find('.'); - if (dotpos != std::string::npos && name.substr(0, dotpos) == "direction") { - try { - uint32_t layer = std::stoi(name.substr(dotpos + 1)); - if (layer == 0) { - fprintf(stderr, "%s: direction tensor invalid in %s\n", __func__, load_info.fname.c_str()); - ggml_free(meta_ctx); - gguf_free(meta_ctx_gguf); - return result; - } - if (layer > max_direction_layer) { - max_direction_layer = layer; - } - } catch (...) { - fprintf(stderr, "%s: direction tensor invalid in %s\n", __func__, load_info.fname.c_str()); - ggml_free(meta_ctx); - gguf_free(meta_ctx_gguf); - return result; - } - } - - struct ggml_tensor * tensor_meta = ggml_get_tensor(meta_ctx, name.c_str()); - if (tensor_meta->type != GGML_TYPE_F32 || ggml_n_dims(tensor_meta) != 1) { - fprintf(stderr, "%s: direction tensor invalid in %s\n", __func__, load_info.fname.c_str()); - ggml_free(meta_ctx); - gguf_free(meta_ctx_gguf); - return result; - } - if (result.n_embd == -1) { - result.n_embd = ggml_nelements(tensor_meta); - } else if (ggml_nelements(tensor_meta) != result.n_embd) { - fprintf(stderr, "%s: direction tensor sizes mismatched in %s\n", __func__, load_info.fname.c_str()); - ggml_free(meta_ctx); - gguf_free(meta_ctx_gguf); - return result; - } - n_bytes += ggml_nbytes(tensor_meta); - } - ggml_free(meta_ctx); - gguf_free(meta_ctx_gguf); + ggml_context * ctx = nullptr; + struct gguf_init_params meta_gguf_params = { + /* .no_alloc = */ false, + /* .ctx = */ &ctx, + }; + struct gguf_context * ctx_gguf = gguf_init_from_file(load_info.fname.c_str(), meta_gguf_params); + if (!ctx_gguf) { + fprintf(stderr, "%s: failed to load control vector file from %s\n", __func__, load_info.fname.c_str()); + return result; } + int32_t n_tensors = gguf_get_n_tensors(ctx_gguf); if (n_tensors == 0) { fprintf(stderr, "%s: no direction tensors found in %s\n", __func__, load_info.fname.c_str()); - return result; } - // load and scale tensors into final control vector context - struct ggml_init_params ggml_params = { - /* .mem_size = */ ggml_tensor_overhead() * n_tensors + n_bytes, - /* .mem_buffer = */ nullptr, - /* .no_alloc = */ false, - }; - struct ggml_context * ctx = ggml_init(ggml_params); + for (int i = 0; i < n_tensors; i++) { + std::string name = gguf_get_tensor_name(ctx_gguf, i); - struct gguf_init_params params = { - /*.no_alloc = */ false, - /*.ctx = */ &ctx, - }; - struct gguf_context * ctx_gguf = gguf_init_from_file(load_info.fname.c_str(), params); - if (!ctx_gguf) { - fprintf(stderr, "%s: failed to load control vector from %s\n", __func__, load_info.fname.c_str()); - ggml_free(ctx); - return result; - } + int layer_idx = -1; - // do not store data for layer 0 (it's not used) - result.data.resize(result.n_embd * max_direction_layer); - - for (uint32_t il = 1; il <= max_direction_layer; il++) { - const std::string name = "direction." + std::to_string(il); - const ggml_tensor * tensor = ggml_get_tensor(ctx, name.c_str()); - - float * dst = result.data.data() + result.n_embd * (il - 1); - - if (tensor) { - const float * src = (const float *) tensor->data; - for (int j = 0; j < result.n_embd; j++) { - dst[j] = src[j] * load_info.strength; - } - } else { - for (int j = 0; j < result.n_embd; j++) { - dst[j] = 0.0f; + // split on '.' + size_t dotpos = name.find('.'); + if (dotpos != std::string::npos && name.substr(0, dotpos) == "direction") { + try { + layer_idx = std::stoi(name.substr(dotpos + 1)); + } catch (...) { + layer_idx = -1; } } + if (layer_idx < 0) { + fprintf(stderr, "%s: invalid/unparsable direction tensor layer index in %s\n", __func__, load_info.fname.c_str()); + result.n_embd = -1; + break; + } else if (layer_idx == 0) { + fprintf(stderr, "%s: invalid (zero) direction tensor layer index in %s\n", __func__, load_info.fname.c_str()); + result.n_embd = -1; + break; + } + + struct ggml_tensor * tensor = ggml_get_tensor(ctx, name.c_str()); + if (tensor->type != GGML_TYPE_F32) { + fprintf(stderr, "%s: invalid (non-F32) direction tensor type in %s\n", __func__, load_info.fname.c_str()); + result.n_embd = -1; + break; + } + if (ggml_n_dims(tensor) != 1) { + fprintf(stderr, "%s: invalid (non-1D) direction tensor shape in %s\n", __func__, load_info.fname.c_str()); + result.n_embd = -1; + break; + } + + if (result.n_embd == -1) { + result.n_embd = ggml_nelements(tensor); + } else if (ggml_nelements(tensor) != result.n_embd) { + fprintf(stderr, "%s: direction tensor in %s does not match previous dimensions\n", __func__, load_info.fname.c_str()); + result.n_embd = -1; + break; + } + + // extend if necessary - do not store data for layer 0 (it's not used) + result.data.resize(std::max(result.data.size(), static_cast(result.n_embd * layer_idx)), 0.0f); + + const float * src = (const float *) tensor->data; + float * dst = result.data.data() + result.n_embd * (layer_idx - 1); // layer 1 at [0] + for (int j = 0; j < result.n_embd; j++) { + dst[j] += src[j] * load_info.strength; // allows multiple directions for same layer in same file + } + } + if (result.n_embd == -1) { + fprintf(stderr, "%s: skipping %s due to invalid direction tensors\n", __func__, load_info.fname.c_str()); + result.data.clear(); + } + + gguf_free(ctx_gguf); + ggml_free(ctx); + return result; } @@ -2933,16 +2984,19 @@ llama_control_vector_data llama_control_vector_load(const std::vector(la).c_str(), std::get<1>(la)); } - fprintf(stream, "lora_base: %s\n", params.lora_base.c_str()); fprintf(stream, "main_gpu: %d # default: 0\n", params.main_gpu); fprintf(stream, "min_keep: %d # default: 0 (disabled)\n", sparams.min_keep); fprintf(stream, "mirostat: %d # default: 0 (disabled)\n", sparams.mirostat); diff --git a/common/common.h b/common/common.h index c541204f6..8240ff99b 100644 --- a/common/common.h +++ b/common/common.h @@ -99,6 +99,7 @@ struct gpt_params { enum llama_split_mode split_mode = LLAMA_SPLIT_MODE_LAYER; // how to split the model across GPUs enum llama_rope_scaling_type rope_scaling_type = LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED; enum llama_pooling_type pooling_type = LLAMA_POOLING_TYPE_UNSPECIFIED; // pooling type for embeddings + enum llama_attention_type attention_type = LLAMA_ATTENTION_TYPE_UNSPECIFIED; // attention type for embeddings // // sampling parameters struct llama_sampling_params sparams; @@ -107,6 +108,7 @@ struct gpt_params { std::string model_draft = ""; // draft model for speculative decoding std::string model_alias = "unknown"; // model alias std::string model_url = ""; // model url to download + std::string hf_token = ""; // HF token std::string hf_repo = ""; // HF repo std::string hf_file = ""; // HF file std::string prompt = ""; @@ -126,7 +128,6 @@ struct gpt_params { // TODO: avoid tuple, use struct std::vector> lora_adapter; // lora adapter path with user defined scale - std::string lora_base = ""; // base model path for the lora adapter std::vector control_vectors; // control vector with user defined scale @@ -200,6 +201,7 @@ struct gpt_params { std::string public_path = ""; std::string chat_template = ""; std::string system_prompt = ""; + bool enable_chat_template = true; std::vector api_keys; @@ -250,8 +252,13 @@ struct gpt_params { std::string cvector_outfile = "control_vector.gguf"; std::string cvector_positive_file = "examples/cvector-generator/positive.txt"; std::string cvector_negative_file = "examples/cvector-generator/negative.txt"; + + bool spm_infill = false; // suffix/prefix/middle pattern for infill + + std::string lora_outfile = "ggml-lora-merged-f16.gguf"; }; +void gpt_params_handle_hf_token(gpt_params & params); void gpt_params_handle_model_default(gpt_params & params); bool gpt_params_parse_ex (int argc, char ** argv, gpt_params & params); @@ -307,8 +314,8 @@ std::tuple llama_init_from_gpt_par struct llama_model_params llama_model_params_from_gpt_params (const gpt_params & params); struct llama_context_params llama_context_params_from_gpt_params(const gpt_params & params); -struct llama_model * llama_load_model_from_url(const char * model_url, const char * path_model, const struct llama_model_params & params); -struct llama_model * llama_load_model_from_hf(const char * repo, const char * file, const char * path_model, const struct llama_model_params & params); +struct llama_model * llama_load_model_from_url(const char * model_url, const char * path_model, const char * hf_token, const struct llama_model_params & params); +struct llama_model * llama_load_model_from_hf(const char * repo, const char * file, const char * path_model, const char * hf_token, const struct llama_model_params & params); // Batch utils @@ -346,21 +353,13 @@ std::string llama_token_to_piece( llama_token token, bool special = true); -// TODO: these should be moved in llama.h C-style API under single `llama_detokenize` function -// that takes into account the tokenizer type and decides how to handle the leading space -// // detokenizes a vector of tokens into a string // should work similar to Python's `tokenizer.decode` -// removes the leading space from the first non-BOS token -std::string llama_detokenize_spm( +// optionally renders special/control tokens +std::string llama_detokenize( llama_context * ctx, - const std::vector & tokens); - -// detokenizes a vector of tokens into a string -// should work similar to Python's `tokenizer.decode` -std::string llama_detokenize_bpe( - llama_context * ctx, - const std::vector & tokens); + const std::vector & tokens, + bool special = true); // Uses the value from the model metadata if possible, otherwise // defaults to true when model type is SPM, otherwise false. @@ -380,6 +379,8 @@ struct llama_chat_msg { bool llama_chat_verify_template(const std::string & tmpl); // CPP wrapper for llama_chat_apply_template +// If the built-in template is not supported, we default to chatml +// If the custom "tmpl" is not supported, we throw an error std::string llama_chat_apply_template(const struct llama_model * model, const std::string & tmpl, const std::vector & chat, @@ -454,4 +455,3 @@ void yaml_dump_string_multiline(FILE * stream, const char * prop_name, const cha void yaml_dump_non_result_info( FILE * stream, const gpt_params & params, const llama_context * lctx, const std::string & timestamp, const std::vector & prompt_tokens, const char * model_desc); - diff --git a/common/json-schema-to-grammar.cpp b/common/json-schema-to-grammar.cpp index 2f233e2e7..881eb49e3 100644 --- a/common/json-schema-to-grammar.cpp +++ b/common/json-schema-to-grammar.cpp @@ -316,7 +316,7 @@ std::unordered_map GRAMMAR_LITERAL_ESCAPES = { }; std::unordered_set NON_LITERAL_SET = {'|', '.', '(', ')', '[', ']', '{', '}', '*', '+', '?'}; -std::unordered_set ESCAPED_IN_REGEXPS_BUT_NOT_IN_LITERALS = {'[', ']', '(', ')', '|', '{', '}', '*', '+', '?'}; +std::unordered_set ESCAPED_IN_REGEXPS_BUT_NOT_IN_LITERALS = {'^', '$', '.', '[', ']', '(', ')', '|', '{', '}', '*', '+', '?'}; template std::string join(Iterator begin, Iterator end, const std::string & separator) { @@ -720,7 +720,7 @@ private: } prop_names.push_back(prop_name); } - if (!(additional_properties.is_boolean() && !additional_properties.get())) { + if ((additional_properties.is_boolean() && additional_properties.get()) || additional_properties.is_object()) { std::string sub_name = name + (name.empty() ? "" : "-") + "additional"; std::string value_rule = additional_properties.is_object() ? visit(additional_properties, sub_name + "-value") diff --git a/common/log.h b/common/log.h index 09fa63c26..1bc5328ce 100644 --- a/common/log.h +++ b/common/log.h @@ -630,7 +630,7 @@ inline std::string LOG_TOKENS_TOSTR_PRETTY(const C & ctx, const T & tokens) buf << "[ "; bool first = true; - for (const auto &token : tokens) + for (const auto & token : tokens) { if (!first) { buf << ", "; diff --git a/common/ngram-cache.h b/common/ngram-cache.h index e4fa4cbd1..ab4c9b376 100644 --- a/common/ngram-cache.h +++ b/common/ngram-cache.h @@ -37,11 +37,18 @@ struct llama_ngram { } }; +struct llama_token_hash_function { + size_t operator()(const llama_token token) const { + // see https://probablydance.com/2018/06/16/fibonacci-hashing-the-optimization-that-the-world-forgot-or-a-better-alternative-to-integer-modulo/ + return token * 11400714819323198485llu; + } +}; + struct llama_ngram_hash_function { size_t operator()(const llama_ngram & ngram) const { - size_t hash = 0; - for (int i = 0; i < LLAMA_NGRAM_MAX; ++i) { - hash ^= std::hash{}(ngram.tokens[i]); + size_t hash = llama_token_hash_function{}(ngram.tokens[0]); + for (int i = 1; i < LLAMA_NGRAM_MAX; ++i) { + hash ^= llama_token_hash_function{}(ngram.tokens[i]); } return hash; } diff --git a/common/sampling.cpp b/common/sampling.cpp index 9f332fe57..079e40516 100644 --- a/common/sampling.cpp +++ b/common/sampling.cpp @@ -282,8 +282,6 @@ static llama_token llama_sampling_sample_impl( GGML_ASSERT(!original_logits.empty()); } llama_token id = 0; - // Get a pointer to the logits - float * logits = llama_get_logits_ith(ctx_main, idx); if (temp < 0.0) { // greedy sampling, with probs @@ -324,12 +322,15 @@ static llama_token llama_sampling_sample_impl( } if (ctx_sampling->grammar != NULL && !is_resampling) { + // Get a pointer to the logits + float * logits = llama_get_logits_ith(ctx_main, idx); + // Create an array with a single token data element for the sampled id llama_token_data single_token_data = {id, logits[id], 0.0f}; llama_token_data_array single_token_data_array = { &single_token_data, 1, false }; // Apply grammar constraints to the single token - llama_sample_grammar(ctx_main, &single_token_data_array, ctx_sampling->grammar); + llama_grammar_sample(ctx_sampling->grammar, ctx_main, &single_token_data_array); // Check if the token is valid according to the grammar by seeing if its logit has been set to -INFINITY bool is_valid = single_token_data_array.data[0].logit != -INFINITY; @@ -377,7 +378,7 @@ static llama_token_data_array llama_sampling_prepare_impl( if (ctx_sampling->grammar != NULL && !apply_grammar) { GGML_ASSERT(original_logits != NULL); // Only make a copy of the original logits if we are not applying grammar checks, not sure if I actually have to do this. - *original_logits = {logits, logits + llama_n_vocab(llama_get_model(ctx_main))}; + *original_logits = {logits, logits + n_vocab}; } // apply params.logit_bias map @@ -390,10 +391,10 @@ static llama_token_data_array llama_sampling_prepare_impl( llama_sample_apply_guidance(ctx_main, logits, logits_guidance, params.cfg_scale); } - cur.clear(); + cur.resize(n_vocab); for (llama_token token_id = 0; token_id < n_vocab; token_id++) { - cur.emplace_back(llama_token_data{token_id, logits[token_id], 0.0f}); + cur[token_id] = llama_token_data{token_id, logits[token_id], 0.0f}; } llama_token_data_array cur_p = { cur.data(), cur.size(), false }; @@ -420,7 +421,7 @@ static llama_token_data_array llama_sampling_prepare_impl( // apply grammar checks before sampling logic if (apply_grammar && ctx_sampling->grammar != NULL) { - llama_sample_grammar(ctx_main, &cur_p, ctx_sampling->grammar); + llama_grammar_sample(ctx_sampling->grammar, ctx_main, &cur_p); } return cur_p; @@ -454,6 +455,6 @@ void llama_sampling_accept( ctx_sampling->prev.push_back(id); if (ctx_sampling->grammar != NULL && apply_grammar) { - llama_grammar_accept_token(ctx_main, ctx_sampling->grammar, id); + llama_grammar_accept_token(ctx_sampling->grammar, ctx_main, id); } } diff --git a/convert-hf-to-gguf.py b/convert_hf_to_gguf.py similarity index 76% rename from convert-hf-to-gguf.py rename to convert_hf_to_gguf.py index eb5aaebac..e93c3829a 100755 --- a/convert-hf-to-gguf.py +++ b/convert_hf_to_gguf.py @@ -13,7 +13,7 @@ import sys from enum import IntEnum from pathlib import Path from hashlib import sha256 -from typing import TYPE_CHECKING, Any, Callable, ContextManager, Iterable, Iterator, Sequence, TypeVar, cast +from typing import TYPE_CHECKING, Any, Callable, ContextManager, Iterable, Iterator, Literal, Sequence, TypeVar, cast import math import numpy as np @@ -48,34 +48,39 @@ class Model: dir_model: Path ftype: gguf.LlamaFileType + fname_out: Path is_big_endian: bool endianess: gguf.GGUFEndian use_temp_file: bool lazy: bool - model_name: str | None part_names: list[str] is_safetensors: bool hparams: dict[str, Any] block_count: int tensor_map: gguf.TensorNameMap tensor_names: set[str] | None - fname_out: Path gguf_writer: gguf.GGUFWriter + model_name: str | None + metadata_override: Path | None + dir_model_card: Path # subclasses should define this! model_arch: gguf.MODEL_ARCH - def __init__(self, dir_model: Path, ftype: gguf.LlamaFileType, fname_out: Path, is_big_endian: bool, use_temp_file: bool, eager: bool, - model_name: str | None, split_max_tensors: int = 0, split_max_size: int = 0, dry_run: bool = False, small_first_shard: bool = False): + def __init__(self, dir_model: Path, ftype: gguf.LlamaFileType, fname_out: Path, is_big_endian: bool = False, + use_temp_file: bool = False, eager: bool = False, + metadata_override: Path | None = None, model_name: str | None = None, + split_max_tensors: int = 0, split_max_size: int = 0, dry_run: bool = False, small_first_shard: bool = False): if type(self) is Model: raise TypeError(f"{type(self).__name__!r} should not be directly instantiated") + self.dir_model = dir_model self.ftype = ftype + self.fname_out = fname_out self.is_big_endian = is_big_endian self.endianess = gguf.GGUFEndian.BIG if is_big_endian else gguf.GGUFEndian.LITTLE self.use_temp_file = use_temp_file self.lazy = not eager - self.model_name = model_name self.part_names = Model.get_model_part_names(self.dir_model, "model", ".safetensors") self.is_safetensors = len(self.part_names) > 0 if not self.is_safetensors: @@ -84,6 +89,11 @@ class Model: self.block_count = self.find_hparam(["n_layers", "num_hidden_layers", "n_layer", "num_layers"]) self.tensor_map = gguf.get_tensor_name_map(self.model_arch, self.block_count) self.tensor_names = None + self.metadata_override = metadata_override + self.model_name = model_name + self.dir_model_card = dir_model # overridden in convert_lora_to_gguf.py + + # Apply heuristics to figure out typical tensor encoding based on first layer tensor encoding type if self.ftype == gguf.LlamaFileType.GUESSED: # NOTE: can't use field "torch_dtype" in config.json, because some finetunes lie. _, first_tensor = next(self.get_tensors()) @@ -93,10 +103,8 @@ class Model: else: logger.info(f"choosing --outtype bf16 from first tensor type ({first_tensor.dtype})") self.ftype = gguf.LlamaFileType.MOSTLY_BF16 - ftype_up: str = self.ftype.name.partition("_")[2].upper() - ftype_lw: str = ftype_up.lower() - # allow templating the file name with the output ftype, useful with the "auto" ftype - self.fname_out = fname_out.parent / fname_out.name.format(ftype_lw, outtype=ftype_lw, ftype=ftype_lw, OUTTYPE=ftype_up, FTYPE=ftype_up) + + # Configure GGUF Writer self.gguf_writer = gguf.GGUFWriter(path=None, arch=gguf.MODEL_ARCH_NAMES[self.model_arch], endianess=self.endianess, use_temp_file=self.use_temp_file, split_max_tensors=split_max_tensors, split_max_size=split_max_size, dry_run=dry_run, small_first_shard=small_first_shard) @@ -148,9 +156,16 @@ class Model: tensor_names_from_parts.update(model_part.keys()) for name in model_part.keys(): - data = model_part.get_tensor(name) if self.is_safetensors else model_part[name] - if self.lazy: - data = LazyTorchTensor.from_eager(data) + if self.is_safetensors: + if self.lazy: + data = model_part.get_slice(name) + data = LazyTorchTensor.from_safetensors_slice(data) + else: + data = model_part.get_tensor(name) + else: + data = model_part[name] + if self.lazy: + data = LazyTorchTensor.from_eager(data) yield name, data # only verify tensor name presence; it doesn't matter if they are not in the right files @@ -186,7 +201,6 @@ class Model: return new_name def set_gguf_parameters(self): - self.gguf_writer.add_name(self.dir_model.name if self.model_name is None else self.model_name) self.gguf_writer.add_block_count(self.block_count) if (n_ctx := self.find_hparam(["max_position_embeddings", "n_ctx"], optional=True)) is not None: @@ -225,6 +239,10 @@ class Model: self.gguf_writer.add_expert_used_count(n_experts_used) logger.info(f"gguf: experts used count = {n_experts_used}") + if (head_dim := self.hparams.get("head_dim")) is not None: + self.gguf_writer.add_key_length(head_dim) + self.gguf_writer.add_value_length(head_dim) + self.gguf_writer.add_file_type(self.ftype) logger.info(f"gguf: file type = {self.ftype}") @@ -243,7 +261,7 @@ class Model: return False - def write_tensors(self): + def prepare_tensors(self): max_name_len = max(len(s) for _, s in self.tensor_map.mapping.values()) + len(".weight,") for name, data_torch in self.get_tensors(): @@ -350,9 +368,62 @@ class Model: self.gguf_writer.add_tensor(new_name, data, raw_dtype=data_qtype) + def set_type(self): + self.gguf_writer.add_type(gguf.GGUFType.MODEL) + + def prepare_metadata(self, vocab_only: bool): + + total_params, shared_params, expert_params, expert_count = self.gguf_writer.get_total_parameter_count() + + self.metadata = gguf.Metadata.load(self.metadata_override, self.dir_model_card, self.model_name, total_params) + + # Fallback to model directory name if metadata name is still missing + if self.metadata.name is None: + self.metadata.name = self.dir_model.name + + # Generate parameter weight class (useful for leader boards) if not yet determined + if self.metadata.size_label is None and total_params > 0: + self.metadata.size_label = gguf.size_label(total_params, shared_params, expert_params, expert_count) + + # Extract the encoding scheme from the file type name. e.g. 'gguf.LlamaFileType.MOSTLY_Q8_0' --> 'Q8_0' + output_type: str = self.ftype.name.partition("_")[2] + + # Filename Output + if self.fname_out.is_dir(): + # Generate default filename based on model specification and available metadata + if not vocab_only: + fname_default: str = gguf.naming_convention(self.metadata.name, self.metadata.basename, self.metadata.finetune, self.metadata.version, self.metadata.size_label, output_type, model_type="LoRA" if total_params < 0 else None) + else: + fname_default: str = gguf.naming_convention(self.metadata.name, self.metadata.basename, self.metadata.finetune, self.metadata.version, size_label=None, output_type=None, model_type="vocab") + + # Use the default filename + self.fname_out = self.fname_out / f"{fname_default}.gguf" + else: + # Output path is a custom defined templated filename + # Note: `not is_dir()` is used because `.is_file()` will not detect + # file template strings as it doesn't actually exist as a file + + # Process templated file name with the output ftype, useful with the "auto" ftype + self.fname_out = self.fname_out.parent / gguf.fill_templated_filename(self.fname_out.name, output_type) + + self.set_type() + + logger.info("Set meta model") + self.metadata.set_gguf_meta_model(self.gguf_writer) + + logger.info("Set model parameters") + self.set_gguf_parameters() + + logger.info("Set model tokenizer") + self.set_vocab() + + logger.info("Set model quantization version") + self.gguf_writer.add_quantization_version(gguf.GGML_QUANT_VERSION) + def write(self): - self.write_tensors() - self.gguf_writer.write_header_to_file(self.fname_out) + self.prepare_tensors() + self.prepare_metadata(vocab_only=False) + self.gguf_writer.write_header_to_file(path=self.fname_out) self.gguf_writer.write_kv_data_to_file() self.gguf_writer.write_tensors_to_file(progress=True) self.gguf_writer.close() @@ -360,7 +431,9 @@ class Model: def write_vocab(self): if len(self.gguf_writer.tensors) != 1: raise ValueError('Splitting the vocabulary is not supported') - self.gguf_writer.write_header_to_file(self.fname_out) + + self.prepare_metadata(vocab_only=True) + self.gguf_writer.write_header_to_file(path=self.fname_out) self.gguf_writer.write_kv_data_to_file() self.gguf_writer.close() @@ -397,6 +470,29 @@ class Model: except KeyError: raise NotImplementedError(f'Architecture {arch!r} not supported!') from None + def does_token_look_special(self, token: str | bytes) -> bool: + if isinstance(token, (bytes, bytearray)): + token_text = token.decode(encoding="utf-8") + elif isinstance(token, memoryview): + token_text = token.tobytes().decode(encoding="utf-8") + else: + token_text = token + + # Some models mark some added tokens which ought to be control tokens as not special. + # (e.g. command-r, command-r-plus, deepseek-coder, gemma{,-2}) + seems_special = token_text in ( + "", # deepseek-coder + "", "<2mass>", "[@BOS@]", # gemma{,-2} + ) + + seems_special = seems_special or (token_text.startswith("<|") and token_text.endswith("|>")) + seems_special = seems_special or (token_text.startswith("<｜") and token_text.endswith("｜>")) # deepseek-coder + + # TODO: should these be marked as UNUSED instead? (maybe not) + seems_special = seems_special or (token_text.startswith("")) # gemma{,-2} + + return seems_special + # used for GPT-2 BPE and WordPiece vocabs def get_vocab_base(self) -> tuple[list[str], list[int], str]: tokens: list[str] = [] @@ -415,20 +511,22 @@ class Model: for i in range(vocab_size): if i not in reverse_vocab: tokens.append(f"[PAD{i}]") - toktypes.append(gguf.TokenType.USER_DEFINED) - elif reverse_vocab[i] in added_vocab: - tokens.append(reverse_vocab[i]) - if tokenizer.added_tokens_decoder[i].special: - toktypes.append(gguf.TokenType.CONTROL) - else: - toktypes.append(gguf.TokenType.USER_DEFINED) + toktypes.append(gguf.TokenType.UNUSED) else: - tokens.append(reverse_vocab[i]) - toktypes.append(gguf.TokenType.NORMAL) + token: str = reverse_vocab[i] + if token in added_vocab: + if tokenizer.added_tokens_decoder[i].special or self.does_token_look_special(token): + toktypes.append(gguf.TokenType.CONTROL) + else: + token = token.replace(b"\xe2\x96\x81".decode("utf-8"), " ") # pre-normalize user-defined spaces + toktypes.append(gguf.TokenType.USER_DEFINED) + else: + toktypes.append(gguf.TokenType.NORMAL) + tokens.append(token) return tokens, toktypes, tokpre - # NOTE: this function is generated by convert-hf-to-gguf-update.py + # NOTE: this function is generated by convert_hf_to_gguf_update.py # do not modify it manually! # ref: https://github.com/ggerganov/llama.cpp/pull/6920 # Marker: Start get_vocab_base_pre @@ -448,7 +546,7 @@ class Model: res = None - # NOTE: if you get an error here, you need to update the convert-hf-to-gguf-update.py script + # NOTE: if you get an error here, you need to update the convert_hf_to_gguf_update.py script # or pull the latest version of the model from Huggingface # don't edit the hashes manually! if chkhsh == "0ef9807a4087ebef797fc749390439009c3b9eda9ad1a097abbe738f486c01e5": @@ -511,15 +609,33 @@ class Model: if chkhsh == "7967bfa498ade6b757b064f31e964dddbb80f8f9a4d68d4ba7998fcf281c531a": # ref: https://huggingface.co/jinaai/jina-embeddings-v2-base-code res = "jina-v2-code" + if chkhsh == "b6e8e1518dc4305be2fe39c313ed643381c4da5db34a98f6a04c093f8afbe99b": + # ref: https://huggingface.co/THUDM/glm-4-9b-chat + res = "chatglm-bpe" + if chkhsh == "7fc505bd3104ca1083b150b17d088b59534ede9bde81f0dd2090967d7fe52cee": + # ref: https://huggingface.co/LumiOpen/Viking-7B + res = "viking" + if chkhsh == "b53802fb28e26d645c3a310b34bfe07da813026ec7c7716883404d5e0f8b1901": + # ref: https://huggingface.co/core42/jais-13b + res = "jais" + if chkhsh == "7b3e7548e4308f52a76e8229e4e6cc831195d0d1df43aed21ac6c93da05fec5f": + # ref: https://huggingface.co/WisdomShell/CodeShell-7B + res = "codeshell" + if chkhsh == "63b97e4253352e6f357cc59ea5b583e3a680eaeaf2632188c2b952de2588485e": + # ref: https://huggingface.co/mistralai/Mistral-Nemo-Base-2407 + res = "tekken" + if chkhsh == "855059429035d75a914d1eda9f10a876752e281a054a7a3d421ef0533e5b6249": + # ref: https://huggingface.co/HuggingFaceTB/SmolLM-135M + res = "smollm" if res is None: logger.warning("\n") logger.warning("**************************************************************************************") logger.warning("** WARNING: The BPE pre-tokenizer was not recognized!") logger.warning("** There are 2 possible reasons for this:") - logger.warning("** - the model has not been added to convert-hf-to-gguf-update.py yet") + logger.warning("** - the model has not been added to convert_hf_to_gguf_update.py yet") logger.warning("** - the pre-tokenization config has changed upstream") - logger.warning("** Check your model files and convert-hf-to-gguf-update.py and update them accordingly.") + logger.warning("** Check your model files and convert_hf_to_gguf_update.py and update them accordingly.") logger.warning("** ref: https://github.com/ggerganov/llama.cpp/pull/6920") logger.warning("**") logger.warning(f"** chkhsh: {chkhsh}") @@ -574,7 +690,7 @@ class Model: for i in range(vocab_size): if i not in reverse_vocab: tokens.append(f"[PAD{i}]") - toktypes.append(gguf.TokenType.USER_DEFINED) + toktypes.append(gguf.TokenType.UNUSED) elif reverse_vocab[i] in added_vocab: tokens.append(reverse_vocab[i]) toktypes.append(gguf.TokenType.CONTROL) @@ -597,15 +713,23 @@ class Model: special_vocab._set_special_token("unk", tokenizer.special_tokens["<|endoftext|>"]) special_vocab.add_to_gguf(self.gguf_writer) - def _set_vocab_sentencepiece(self): + def _set_vocab_sentencepiece(self, add_to_gguf=True): + tokens, scores, toktypes = self._create_vocab_sentencepiece() + + self.gguf_writer.add_tokenizer_model("llama") + self.gguf_writer.add_tokenizer_pre("default") + self.gguf_writer.add_token_list(tokens) + self.gguf_writer.add_token_scores(scores) + self.gguf_writer.add_token_types(toktypes) + + special_vocab = gguf.SpecialVocab(self.dir_model, n_vocab=len(tokens)) + special_vocab.add_to_gguf(self.gguf_writer) + + def _create_vocab_sentencepiece(self): from sentencepiece import SentencePieceProcessor tokenizer_path = self.dir_model / 'tokenizer.model' - tokens: list[bytes] = [] - scores: list[float] = [] - toktypes: list[int] = [] - if not tokenizer_path.is_file(): raise FileNotFoundError(f"File not found: {tokenizer_path}") @@ -616,7 +740,7 @@ class Model: tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)] scores: list[float] = [-10000.0] * vocab_size - toktypes: list[int] = [SentencePieceTokenTypes.UNKNOWN] * vocab_size + toktypes: list[int] = [SentencePieceTokenTypes.UNUSED] * vocab_size for token_id in range(tokenizer.vocab_size()): piece = tokenizer.IdToPiece(token_id) @@ -643,7 +767,7 @@ class Model: added_tokens_json = json.load(f) for key in added_tokens_json: token_id = added_tokens_json[key] - if (token_id >= vocab_size): + if token_id >= vocab_size: logger.warning(f'ignore token {token_id}: id is out of range, max={vocab_size - 1}') continue @@ -651,6 +775,26 @@ class Model: scores[token_id] = -1000.0 toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED + tokenizer_config_file = self.dir_model / 'tokenizer_config.json' + if tokenizer_config_file.is_file(): + with open(tokenizer_config_file, "r", encoding="utf-8") as f: + tokenizer_config_json = json.load(f) + added_tokens_decoder = tokenizer_config_json.get("added_tokens_decoder", {}) + for token_id, token_data in added_tokens_decoder.items(): + token_id = int(token_id) + token: str = token_data["content"] + if toktypes[token_id] != SentencePieceTokenTypes.UNUSED: + if tokens[token_id] != token.encode("utf-8"): + logger.warning(f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r} -> {token!r}') + if token_data.get("special") or self.does_token_look_special(token): + toktypes[token_id] = SentencePieceTokenTypes.CONTROL + else: + token = token.replace(b"\xe2\x96\x81".decode("utf-8"), " ") # pre-normalize user-defined spaces + toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED + + scores[token_id] = -1000.0 + tokens[token_id] = token.encode("utf-8") + if vocab_size > len(tokens): pad_count = vocab_size - len(tokens) logger.debug(f"Padding vocab with {pad_count} token(s) - [PAD1] through [PAD{pad_count}]") @@ -659,14 +803,7 @@ class Model: scores.append(-1000.0) toktypes.append(SentencePieceTokenTypes.UNUSED) - self.gguf_writer.add_tokenizer_model("llama") - self.gguf_writer.add_tokenizer_pre("default") - self.gguf_writer.add_token_list(tokens) - self.gguf_writer.add_token_scores(scores) - self.gguf_writer.add_token_types(toktypes) - - special_vocab = gguf.SpecialVocab(self.dir_model, n_vocab=len(tokens)) - special_vocab.add_to_gguf(self.gguf_writer) + return tokens, scores, toktypes def _set_vocab_llama_hf(self): vocab = gguf.LlamaHfVocab(self.dir_model) @@ -690,6 +827,51 @@ class Model: special_vocab = gguf.SpecialVocab(self.dir_model, n_vocab=len(tokens)) special_vocab.add_to_gguf(self.gguf_writer) + def _set_vocab_builtin(self, model_name: Literal["gpt-neox", "llama-spm"], vocab_size: int): + tokenizer_path = Path(sys.path[0]) / "models" / f"ggml-vocab-{model_name}.gguf" + logger.warning(f"Using tokenizer from '{os.path.relpath(tokenizer_path, os.getcwd())}'") + vocab_reader = gguf.GGUFReader(tokenizer_path, "r") + + default_pre = "mpt" if model_name == "gpt-neox" else "default" + + field = vocab_reader.get_field(gguf.Keys.Tokenizer.MODEL) + assert field # tokenizer model + self.gguf_writer.add_tokenizer_model(bytes(field.parts[-1]).decode("utf-8")) + + field = vocab_reader.get_field(gguf.Keys.Tokenizer.PRE) + self.gguf_writer.add_tokenizer_pre(bytes(field.parts[-1]).decode("utf-8") if field else default_pre) + + field = vocab_reader.get_field(gguf.Keys.Tokenizer.LIST) + assert field # token list + self.gguf_writer.add_token_list([bytes(field.parts[i]) for i in field.data][:vocab_size]) + + if model_name == "llama-spm": + field = vocab_reader.get_field(gguf.Keys.Tokenizer.SCORES) + assert field # token scores + self.gguf_writer.add_token_scores([field.parts[i].tolist()[0] for i in field.data][:vocab_size]) + + field = vocab_reader.get_field(gguf.Keys.Tokenizer.TOKEN_TYPE) + assert field # token types + self.gguf_writer.add_token_types([field.parts[i].tolist()[0] for i in field.data][:vocab_size]) + + if model_name != "llama-spm": + field = vocab_reader.get_field(gguf.Keys.Tokenizer.MERGES) + assert field # token merges + self.gguf_writer.add_token_merges([bytes(field.parts[i]) for i in field.data]) + + if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.BOS_ID)) is not None: + self.gguf_writer.add_bos_token_id(field.parts[-1].tolist()[0]) + if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.EOS_ID)) is not None: + self.gguf_writer.add_eos_token_id(field.parts[-1].tolist()[0]) + if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.UNK_ID)) is not None: + self.gguf_writer.add_unk_token_id(field.parts[-1].tolist()[0]) + if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.PAD_ID)) is not None: + self.gguf_writer.add_pad_token_id(field.parts[-1].tolist()[0]) + if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.ADD_BOS)) is not None: + self.gguf_writer.add_add_bos_token(field.parts[-1].tolist()[0]) + if (field := vocab_reader.get_field(gguf.Keys.Tokenizer.ADD_EOS)) is not None: + self.gguf_writer.add_add_eos_token(field.parts[-1].tolist()[0]) + @Model.register("GPTNeoXForCausalLM") class GPTNeoXModel(Model): @@ -698,7 +880,6 @@ class GPTNeoXModel(Model): def set_gguf_parameters(self): block_count = self.hparams["num_hidden_layers"] - self.gguf_writer.add_name(self.dir_model.name if self.model_name is None else self.model_name) self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"]) self.gguf_writer.add_embedding_length(self.hparams["hidden_size"]) self.gguf_writer.add_block_count(block_count) @@ -754,7 +935,6 @@ class BloomModel(Model): model_arch = gguf.MODEL_ARCH.BLOOM def set_gguf_parameters(self): - self.gguf_writer.add_name("Bloom") n_embed = self.hparams.get("hidden_size", self.hparams.get("n_embed")) n_head = self.hparams.get("n_head", self.hparams.get("num_attention_heads")) self.gguf_writer.add_context_length(self.hparams.get("seq_length", n_embed)) @@ -831,7 +1011,6 @@ class MPTModel(Model): def set_gguf_parameters(self): block_count = self.hparams["n_layers"] - self.gguf_writer.add_name(self.dir_model.name if self.model_name is None else self.model_name) self.gguf_writer.add_context_length(self.hparams["max_seq_len"]) self.gguf_writer.add_embedding_length(self.hparams["d_model"]) self.gguf_writer.add_block_count(block_count) @@ -870,7 +1049,6 @@ class OrionModel(Model): block_count = self.hparams["num_hidden_layers"] head_count = self.hparams["num_attention_heads"] head_count_kv = self.hparams.get("num_key_value_heads", head_count) - hf_repo = self.hparams.get("_name_or_path", "") ctx_length = 0 if "max_sequence_length" in self.hparams: @@ -883,8 +1061,6 @@ class OrionModel(Model): raise ValueError("gguf: can not find ctx length parameter.") self.gguf_writer.add_file_type(self.ftype) - self.gguf_writer.add_name(self.dir_model.name if self.model_name is None else self.model_name) - self.gguf_writer.add_source_hf_repo(hf_repo) self.gguf_writer.add_tensor_data_layout("Meta AI original pth") self.gguf_writer.add_context_length(ctx_length) self.gguf_writer.add_embedding_length(self.hparams["hidden_size"]) @@ -908,7 +1084,6 @@ class BaichuanModel(Model): block_count = self.hparams["num_hidden_layers"] head_count = self.hparams["num_attention_heads"] head_count_kv = self.hparams.get("num_key_value_heads", head_count) - hf_repo = self.hparams.get("_name_or_path", "") ctx_length = 0 if "max_sequence_length" in self.hparams: @@ -920,8 +1095,6 @@ class BaichuanModel(Model): else: raise ValueError("gguf: can not find ctx length parameter.") - self.gguf_writer.add_name(self.dir_model.name if self.model_name is None else self.model_name) - self.gguf_writer.add_source_hf_repo(hf_repo) self.gguf_writer.add_tensor_data_layout("Meta AI original pth") self.gguf_writer.add_context_length(ctx_length) self.gguf_writer.add_embedding_length(self.hparams["hidden_size"]) @@ -1035,7 +1208,6 @@ class XverseModel(Model): block_count = self.hparams["num_hidden_layers"] head_count = self.hparams["num_attention_heads"] head_count_kv = self.hparams.get("num_key_value_heads", head_count) - hf_repo = self.hparams.get("_name_or_path", "") ctx_length = 0 if "max_sequence_length" in self.hparams: @@ -1047,8 +1219,6 @@ class XverseModel(Model): else: raise ValueError("gguf: can not find ctx length parameter.") - self.gguf_writer.add_name(self.dir_model.name if self.model_name is None else self.model_name) - self.gguf_writer.add_source_hf_repo(hf_repo) self.gguf_writer.add_tensor_data_layout("Meta AI original pth") self.gguf_writer.add_context_length(ctx_length) self.gguf_writer.add_embedding_length(self.hparams["hidden_size"]) @@ -1107,7 +1277,6 @@ class FalconModel(Model): if n_head_kv is None: n_head_kv = self.hparams.get("n_head_kv", 1) # old name - self.gguf_writer.add_name("Falcon") self.gguf_writer.add_context_length(2048) # not in config.json self.gguf_writer.add_tensor_data_layout("jploski") # qkv tensor transform self.gguf_writer.add_embedding_length(self.hparams["hidden_size"]) @@ -1152,7 +1321,6 @@ class StarCoderModel(Model): def set_gguf_parameters(self): block_count = self.hparams["n_layer"] - self.gguf_writer.add_name("StarCoder") self.gguf_writer.add_context_length(self.hparams["n_positions"]) self.gguf_writer.add_embedding_length(self.hparams["n_embd"]) self.gguf_writer.add_feed_forward_length(4 * self.hparams["n_embd"]) @@ -1172,11 +1340,11 @@ class RefactModel(Model): # TODO: how to determine special FIM tokens automatically? special_vocab = gguf.SpecialVocab(self.dir_model, load_merges=False, - special_token_types = ['prefix', 'suffix', 'middle', 'fsep', 'eot']) + special_token_types = ['prefix', 'suffix', 'middle', 'eot']) special_vocab._set_special_token("prefix", 1) special_vocab._set_special_token("suffix", 3) special_vocab._set_special_token("middle", 2) - special_vocab._set_special_token("fsep", 4) # is this correct? + special_vocab.chat_template = None # do not add it twice special_vocab.add_to_gguf(self.gguf_writer) def set_gguf_parameters(self): @@ -1188,7 +1356,6 @@ class RefactModel(Model): block_count = self.hparams["n_layer"] - self.gguf_writer.add_name("Refact") # refact uses Alibi. So this is from config.json which might be used by training. self.gguf_writer.add_context_length(self.hparams["n_positions"]) self.gguf_writer.add_embedding_length(self.hparams["n_embd"]) @@ -1236,14 +1403,13 @@ class StableLMModel(Model): if (self.dir_model / "tokenizer.json").is_file(): self._set_vocab_gpt2() else: - # StableLM 2 1.6B uses a vocab in a similar format to Qwen's vocab + # StableLM 2 1.6B used to have a vocab in a similar format to Qwen's vocab self._set_vocab_qwen() def set_gguf_parameters(self): hparams = self.hparams block_count = hparams["num_hidden_layers"] - self.gguf_writer.add_name(self.dir_model.name if self.model_name is None else self.model_name) self.gguf_writer.add_context_length(hparams["max_position_embeddings"]) self.gguf_writer.add_embedding_length(hparams["hidden_size"]) self.gguf_writer.add_block_count(block_count) @@ -1305,8 +1471,8 @@ class StableLMModel(Model): return [(new_name, data_torch)] - def write_tensors(self): - super().write_tensors() + def prepare_tensors(self): + super().prepare_tensors() if self._q_norms is not None or self._k_norms is not None: # flatten two `list[dict[str, Tensor]]` into a single `list[str]` @@ -1325,7 +1491,7 @@ class LlamaModel(Model): def set_vocab(self): try: - self. _set_vocab_sentencepiece() + self._set_vocab_sentencepiece() except FileNotFoundError: try: self._set_vocab_llama_hf() @@ -1349,7 +1515,12 @@ class LlamaModel(Model): super().set_gguf_parameters() hparams = self.hparams self.gguf_writer.add_vocab_size(hparams["vocab_size"]) - self.gguf_writer.add_rope_dimension_count(hparams["hidden_size"] // hparams["num_attention_heads"]) + + if "head_dim" in hparams: + rope_dim = hparams["head_dim"] + else: + rope_dim = hparams["hidden_size"] // hparams["num_attention_heads"] + self.gguf_writer.add_rope_dimension_count(rope_dim) if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]: if self.hparams["rope_scaling"].get("type") == "linear": @@ -1422,8 +1593,36 @@ class LlamaModel(Model): return [(self.map_tensor_name(name), data_torch)] - def write_tensors(self): - super().write_tensors() + def prepare_tensors(self): + if rope_scaling := self.find_hparam(["rope_scaling"], optional=True): + if rope_scaling.get("rope_type", '').lower() == "llama3": + base = self.hparams.get("rope_theta", 10000.0) + dim = self.hparams["hidden_size"] // self.hparams["num_attention_heads"] + freqs = 1.0 / (base ** (torch.arange(0, dim, 2, dtype=torch.float32) / dim)) + + factor = rope_scaling.get("factor", 8.0) + low_freq_factor = rope_scaling.get("low_freq_factor", 1.0) + high_freq_factor = rope_scaling.get("high_freq_factor", 4.0) + old_context_len = self.hparams.get("original_max_position_embeddings", 8192) + + low_freq_wavelen = old_context_len / low_freq_factor + high_freq_wavelen = old_context_len / high_freq_factor + assert low_freq_wavelen != high_freq_wavelen + + rope_factors = [] + for freq in freqs: + wavelen = 2 * math.pi / freq + if wavelen < high_freq_wavelen: + rope_factors.append(1) + elif wavelen > low_freq_wavelen: + rope_factors.append(factor) + else: + smooth = (old_context_len / wavelen - low_freq_factor) / (high_freq_factor - low_freq_factor) + rope_factors.append(1 / ((1 - smooth) / factor + smooth)) + + self.gguf_writer.add_tensor(self.format_tensor_name(gguf.MODEL_TENSOR.ROPE_FREQS), np.array(rope_factors, dtype=np.float32)) + + super().prepare_tensors() if self._experts is not None: # flatten `list[dict[str, Tensor]]` into `list[str]` @@ -1493,7 +1692,6 @@ class GrokModel(Model): def set_gguf_parameters(self): super().set_gguf_parameters() - self.gguf_writer.add_name("Grok") _experts: list[dict[str, Tensor]] | None = None @@ -1542,7 +1740,6 @@ class DbrxModel(Model): def set_gguf_parameters(self): ffn_config = self.hparams["ffn_config"] attn_config = self.hparams["attn_config"] - self.gguf_writer.add_name(self.hparams["model_type"]) self.gguf_writer.add_block_count(self.hparams["n_layers"]) self.gguf_writer.add_context_length(self.hparams["max_seq_len"]) @@ -1555,7 +1752,6 @@ class DbrxModel(Model): self.gguf_writer.add_rope_freq_base(attn_config["rope_theta"]) self.gguf_writer.add_clamp_kqv(attn_config["clip_qkv"]) - self.gguf_writer.add_file_type(self.ftype) self.gguf_writer.add_expert_count(ffn_config["moe_num_experts"]) self.gguf_writer.add_expert_used_count(ffn_config["moe_top_k"]) @@ -1612,7 +1808,6 @@ class MiniCPMModel(Model): def set_gguf_parameters(self): block_count = self.hparams["num_hidden_layers"] - self.gguf_writer.add_name("MiniCPM") self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"]) self.gguf_writer.add_embedding_length(self.hparams["hidden_size"]) self.gguf_writer.add_block_count(block_count) @@ -1682,7 +1877,6 @@ class QwenModel(Model): self._set_vocab_qwen() def set_gguf_parameters(self): - self.gguf_writer.add_name("Qwen") self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"]) self.gguf_writer.add_block_count(self.hparams["num_hidden_layers"]) self.gguf_writer.add_embedding_length(self.hparams["hidden_size"]) @@ -1758,8 +1952,8 @@ class Qwen2MoeModel(Model): return [(self.map_tensor_name(name), data_torch)] - def write_tensors(self): - super().write_tensors() + def prepare_tensors(self): + super().prepare_tensors() if self._experts is not None: # flatten `list[dict[str, Tensor]]` into `list[str]` @@ -1773,7 +1967,6 @@ class GPT2Model(Model): model_arch = gguf.MODEL_ARCH.GPT2 def set_gguf_parameters(self): - self.gguf_writer.add_name(self.dir_model.name if self.model_name is None else self.model_name) self.gguf_writer.add_block_count(self.hparams["n_layer"]) self.gguf_writer.add_context_length(self.hparams["n_ctx"]) self.gguf_writer.add_embedding_length(self.hparams["n_embd"]) @@ -1816,7 +2009,6 @@ class Phi2Model(Model): n_embd = self.find_hparam(["hidden_size", "n_embd"]) n_head = self.find_hparam(["num_attention_heads", "n_head"]) - self.gguf_writer.add_name("Phi2") self.gguf_writer.add_context_length(self.find_hparam(["n_positions", "max_position_embeddings"])) self.gguf_writer.add_embedding_length(n_embd) @@ -1849,7 +2041,7 @@ class Phi3MiniModel(Model): tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)] scores: list[float] = [-10000.0] * vocab_size - toktypes: list[int] = [SentencePieceTokenTypes.UNKNOWN] * vocab_size + toktypes: list[int] = [SentencePieceTokenTypes.UNUSED] * vocab_size for token_id in range(tokenizer.vocab_size()): @@ -1878,7 +2070,7 @@ class Phi3MiniModel(Model): for key in added_tokens_json: token_id = added_tokens_json[key] - if (token_id >= vocab_size): + if token_id >= vocab_size: logger.debug(f'ignore token {token_id}: id is out of range, max={vocab_size - 1}') continue @@ -1894,8 +2086,9 @@ class Phi3MiniModel(Model): for token_id, foken_data in added_tokens_decoder.items(): token_id = int(token_id) token = foken_data["content"].encode("utf-8") - if toktypes[token_id] != SentencePieceTokenTypes.UNKNOWN: - assert tokens[token_id] == token + if toktypes[token_id] != SentencePieceTokenTypes.UNUSED: + if tokens[token_id] != token: + logger.warning(f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r} -> {token.decode("utf-8")!r}') tokens[token_id] = token scores[token_id] = -1000.0 toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED @@ -1910,8 +2103,9 @@ class Phi3MiniModel(Model): for foken_data in added_tokens: token_id = int(foken_data["id"]) token = foken_data["content"].encode("utf-8") - if toktypes[token_id] != SentencePieceTokenTypes.UNKNOWN: - assert tokens[token_id] == token + if toktypes[token_id] != SentencePieceTokenTypes.UNUSED: + if tokens[token_id] != token: + logger.warning(f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r} -> {token.decode("utf-8")!r}') tokens[token_id] = token scores[token_id] = -1000.0 toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED @@ -1938,7 +2132,6 @@ class Phi3MiniModel(Model): orig_max_pos_embds = self.find_hparam(["original_max_position_embeddings"]) rope_dims = n_embd // n_head - self.gguf_writer.add_name("Phi3") self.gguf_writer.add_context_length(max_pos_embds) self.gguf_writer.add_rope_scaling_orig_ctx_len(orig_max_pos_embds) self.gguf_writer.add_embedding_length(n_embd) @@ -1950,10 +2143,11 @@ class Phi3MiniModel(Model): self.gguf_writer.add_rope_dimension_count(rope_dims) self.gguf_writer.add_rope_freq_base(self.find_hparam(["rope_theta"])) self.gguf_writer.add_file_type(self.ftype) + self.gguf_writer.add_sliding_window(self.find_hparam(["sliding_window"])) # write rope scaling for long context (128k) model rope_scaling = self.find_hparam(['rope_scaling'], True) - if (rope_scaling is None): + if rope_scaling is None: return scale = max_pos_embds / orig_max_pos_embds @@ -1962,7 +2156,7 @@ class Phi3MiniModel(Model): if len(rope_scaling_type) == 0: raise KeyError('Missing the required key rope_scaling.type') - if rope_scaling_type == 'su': + if rope_scaling_type == 'su' or rope_scaling_type == 'longrope': attn_factor = math.sqrt(1 + math.log(scale) / math.log(orig_max_pos_embds)) if scale > 1.0 else 1.0 elif rope_scaling_type == 'yarn': attn_factor = 0.1 * math.log(scale) + 1.0 if scale > 1.0 else 1.0 @@ -1995,7 +2189,6 @@ class PlamoModel(Model): hparams = self.hparams block_count = hparams["num_hidden_layers"] - self.gguf_writer.add_name("PLaMo") self.gguf_writer.add_context_length(4096) # not in config.json self.gguf_writer.add_embedding_length(hparams["hidden_size"]) self.gguf_writer.add_feed_forward_length(hparams["intermediate_size"]) @@ -2040,7 +2233,6 @@ class CodeShellModel(Model): def set_gguf_parameters(self): block_count = self.hparams["n_layer"] - self.gguf_writer.add_name("CodeShell") self.gguf_writer.add_context_length(self.hparams["n_positions"]) self.gguf_writer.add_embedding_length(self.hparams["n_embd"]) self.gguf_writer.add_feed_forward_length(4 * self.hparams["n_embd"]) @@ -2092,7 +2284,7 @@ class InternLM2Model(Model): logger.error(f'Error: Missing {tokenizer_path}') sys.exit(1) - sentencepiece_model = model.ModelProto() + sentencepiece_model = model.ModelProto() # pyright: ignore[reportAttributeAccessIssue] sentencepiece_model.ParseFromString(open(tokenizer_path, "rb").read()) add_prefix = sentencepiece_model.normalizer_spec.add_dummy_prefix @@ -2120,6 +2312,9 @@ class InternLM2Model(Model): toktype = SentencePieceTokenTypes.UNUSED elif tokenizer.IsByte(token_id): toktype = SentencePieceTokenTypes.BYTE + # take care of ununsed raw token + if piece.startswith('[UNUSED'): + toktype = SentencePieceTokenTypes.UNUSED tokens.append(text) scores.append(score) @@ -2135,6 +2330,49 @@ class InternLM2Model(Model): scores.append(-1000.0) toktypes.append(SentencePieceTokenTypes.USER_DEFINED) + chat_eos_token = '<|im_end|>' + chat_eos_token_id = None + + tokenizer_config_file = self.dir_model / 'tokenizer_config.json' + if tokenizer_config_file.is_file(): + with open(tokenizer_config_file, "r", encoding="utf-8") as f: + tokenizer_config_json = json.load(f) + added_tokens_decoder = tokenizer_config_json.get("added_tokens_decoder", {}) + for token_id, foken_data in added_tokens_decoder.items(): + token_id = int(token_id) + token = foken_data["content"] + if token == chat_eos_token: + chat_eos_token_id = token_id + token = token.encode("utf-8") + if toktypes[token_id] != SentencePieceTokenTypes.UNUSED: + if tokens[token_id] != token: + logger.warning(f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r} -> {token.decode("utf-8")!r}') + tokens[token_id] = token + scores[token_id] = -1000.0 + toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED + if foken_data.get("special"): + toktypes[token_id] = SentencePieceTokenTypes.CONTROL + + tokenizer_file = self.dir_model / 'tokenizer.json' + if tokenizer_file.is_file(): + with open(tokenizer_file, "r", encoding="utf-8") as f: + tokenizer_json = json.load(f) + added_tokens = tokenizer_json.get("added_tokens", []) + for foken_data in added_tokens: + token_id = int(foken_data["id"]) + token = foken_data["content"] + if token == chat_eos_token: + chat_eos_token_id = token_id + token = token.encode("utf-8") + if toktypes[token_id] != SentencePieceTokenTypes.UNUSED: + if tokens[token_id] != token: + logger.warning(f'replacing token {token_id}: {tokens[token_id].decode("utf-8")!r} -> {token.decode("utf-8")!r}') + tokens[token_id] = token + scores[token_id] = -1000.0 + toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED + if foken_data.get("special"): + toktypes[token_id] = SentencePieceTokenTypes.CONTROL + self.gguf_writer.add_tokenizer_model("llama") self.gguf_writer.add_tokenizer_pre("default") self.gguf_writer.add_token_list(tokens) @@ -2144,37 +2382,17 @@ class InternLM2Model(Model): special_vocab = gguf.SpecialVocab(self.dir_model, n_vocab=len(tokens)) old_eos = special_vocab.special_token_ids["eos"] - if "chat" in os.path.basename(self.dir_model.absolute()): + if chat_eos_token_id is not None: # For the chat model, we replace the eos with '<|im_end|>'. # TODO: this is a hack, should be fixed # https://github.com/ggerganov/llama.cpp/pull/6745#issuecomment-2067687048 - special_vocab.special_token_ids["eos"] = self._try_get_sft_eos(tokenizer) - logger.warning(f"Replace eos:{old_eos} with a special token:{special_vocab.special_token_ids['eos']} \ -in chat mode so that the conversation can end normally.") + special_vocab.special_token_ids["eos"] = chat_eos_token_id + logger.warning(f"Replace eos:{old_eos} with a special token:{chat_eos_token_id}" + " in chat mode so that the conversation can end normally.") special_vocab.add_to_gguf(self.gguf_writer) - def _try_get_sft_eos(self, tokenizer): - unused_145_list = tokenizer.Encode('[UNUSED_TOKEN_145]') - im_end_list = tokenizer.Encode('<|im_end|>') - eos_token = None - assert (len(unused_145_list) == 1) ^ (len(im_end_list) == 1) - if len(unused_145_list) == 1: - eos_token = unused_145_list[0] - if len(im_end_list) == 1: - eos_token = im_end_list[0] - assert eos_token - return eos_token - - def _hf_permute_qk(self, weights, n_head: int, n_head_kv: int): - if n_head_kv is not None and n_head != n_head_kv: - n_head = n_head_kv - return (weights.reshape(n_head, 2, weights.shape[0] // n_head // 2, *weights.shape[1:]) - .swapaxes(1, 2) - .reshape(weights.shape)) - def set_gguf_parameters(self): - self.gguf_writer.add_name("InternLM2") self.gguf_writer.add_context_length(self.hparams["max_position_embeddings"]) self.gguf_writer.add_block_count(self.hparams["num_hidden_layers"]) self.gguf_writer.add_embedding_length(self.hparams["hidden_size"]) @@ -2184,30 +2402,30 @@ in chat mode so that the conversation can end normally.") self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"]) self.gguf_writer.add_head_count_kv(self.hparams["num_key_value_heads"]) self.gguf_writer.add_file_type(self.ftype) + if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]: + if self.hparams["rope_scaling"].get("type") == "linear": + self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR) + self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"]) def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: num_heads = self.hparams["num_attention_heads"] num_kv_heads = self.hparams["num_key_value_heads"] - hidden_size = self.hparams["hidden_size"] + n_embd = self.hparams["hidden_size"] q_per_kv = num_heads // num_kv_heads - head_dim = hidden_size // num_heads + head_dim = n_embd // num_heads num_groups = num_heads // q_per_kv - qkv_pattern = r"model\.layers\.(\d+)\.attention\.wqkv" - - if re.match(qkv_pattern, name): - bid = re.findall(qkv_pattern, name)[0] + if bid is not None and f"model.layers.{bid}.attention.wqkv" in name: qkv = data_torch - # qkv = rearrange(qkv.T, " o (g n i) ->o g n i", g=num_groups, n=q_per_kv + 2, i=head_dim) - qkv = qkv.T.reshape((-1, num_groups, q_per_kv + 2, head_dim)) - q, k, v = qkv[..., : q_per_kv, :], qkv[..., q_per_kv: q_per_kv + 1, :], qkv[..., q_per_kv + 1: q_per_kv + 2, :] + + qkv = qkv.reshape((num_groups, q_per_kv + 2, head_dim, n_embd)) + q, k, v = qkv[:, : q_per_kv], qkv[:, -2], qkv[:, -1] + # The model weights of q and k equire additional reshape. - # q = self._hf_permute_qk(rearrange(q, " o g n i -> o (g n i)").T, num_heads, num_heads) - q = self._hf_permute_qk(q.reshape((q.shape[0], -1)).T, num_heads, num_heads) - # k = self._hf_permute_qk(rearrange(k, " o g n i -> o (g n i)").T, num_heads, num_kv_heads) - k = self._hf_permute_qk(k.reshape((k.shape[0], -1)).T, num_heads, num_kv_heads) - # v = rearrange(v, " o g n i -> o (g n i)").T - v = v.reshape((v.shape[0], -1)).T + q = LlamaModel.permute(q.reshape((-1, q.shape[-1])), num_heads, num_heads) + k = LlamaModel.permute(k.reshape((-1, k.shape[-1])), num_heads, num_kv_heads) + v = v.reshape((-1, v.shape[-1])) + return [ (self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_Q, bid), q), (self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_K, bid), k), @@ -2334,13 +2552,15 @@ class GemmaModel(Model): special_vocab._set_special_token("middle", 68) special_vocab._set_special_token("fsep", 70) special_vocab._set_special_token("eot", 107) + special_vocab.chat_template = None # do not add it twice special_vocab.add_to_gguf(self.gguf_writer) + self.gguf_writer.add_add_space_prefix(False) + def set_gguf_parameters(self): hparams = self.hparams block_count = hparams["num_hidden_layers"] - self.gguf_writer.add_name(self.dir_model.name if self.model_name is None else self.model_name) self.gguf_writer.add_context_length(hparams["max_position_embeddings"]) self.gguf_writer.add_embedding_length(hparams["hidden_size"]) self.gguf_writer.add_block_count(block_count) @@ -2368,6 +2588,53 @@ class GemmaModel(Model): return [(self.map_tensor_name(name), data_torch)] +@Model.register("Gemma2ForCausalLM") +class Gemma2Model(Model): + model_arch = gguf.MODEL_ARCH.GEMMA2 + + def set_vocab(self): + self._set_vocab_sentencepiece() + + self.gguf_writer.add_add_space_prefix(False) + + def set_gguf_parameters(self): + hparams = self.hparams + block_count = hparams["num_hidden_layers"] + + self.gguf_writer.add_context_length(hparams["max_position_embeddings"]) + self.gguf_writer.add_embedding_length(hparams["hidden_size"]) + self.gguf_writer.add_block_count(block_count) + self.gguf_writer.add_feed_forward_length(hparams["intermediate_size"]) + self.gguf_writer.add_head_count(hparams["num_attention_heads"]) + self.gguf_writer.add_head_count_kv(self.hparams["num_key_value_heads"] if "num_key_value_heads" in hparams else hparams["num_attention_heads"]) + self.gguf_writer.add_layer_norm_rms_eps(self.hparams["rms_norm_eps"]) + self.gguf_writer.add_key_length(hparams["head_dim"]) + self.gguf_writer.add_value_length(hparams["head_dim"]) + self.gguf_writer.add_file_type(self.ftype) + self.gguf_writer.add_attn_logit_softcapping( + self.hparams["attn_logit_softcapping"] + ) + self.gguf_writer.add_final_logit_softcapping( + self.hparams["final_logit_softcapping"] + ) + self.gguf_writer.add_sliding_window(self.hparams["sliding_window"]) + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + del bid # unused + + # lm_head is not used in llama.cpp, while autoawq will include this tensor in model + # To prevent errors, skip loading lm_head.weight. + if name == "lm_head.weight": + logger.debug(f"Skipping get tensor {name!r} in safetensors so that convert can end normally.") + return [] + + # ref: https://github.com/huggingface/transformers/blob/fc37f38915372c15992b540dfcbbe00a916d4fc6/src/transformers/models/gemma/modeling_gemma.py#L89 + if name.endswith("norm.weight"): + data_torch = data_torch + 1 + + return [(self.map_tensor_name(name), data_torch)] + + @Model.register("Starcoder2ForCausalLM") class StarCoder2Model(Model): model_arch = gguf.MODEL_ARCH.STARCODER2 @@ -2392,39 +2659,7 @@ class MambaModel(Model): self._set_vocab_sentencepiece() else: # Use the GPT-NeoX tokenizer when no tokenizer files are present - tokenizer_path = Path(sys.path[0]) / "models" / "ggml-vocab-gpt-neox.gguf" - logger.warning(f"Using tokenizer from '{os.path.relpath(tokenizer_path, os.getcwd())}'") - neox_reader = gguf.GGUFReader(tokenizer_path, "r") - - field = neox_reader.get_field(gguf.Keys.Tokenizer.MODEL) - self.gguf_writer.add_tokenizer_model(bytes(field.parts[-1]).decode("utf-8") if field else "gpt2") - - field = neox_reader.get_field(gguf.Keys.Tokenizer.PRE) - self.gguf_writer.add_tokenizer_pre(bytes(field.parts[-1]).decode("utf-8") if field else "mpt") - - field = neox_reader.get_field(gguf.Keys.Tokenizer.LIST) - assert field - self.gguf_writer.add_token_list([bytes(field.parts[i]) for i in field.data][:vocab_size]) - - field = neox_reader.get_field(gguf.Keys.Tokenizer.TOKEN_TYPE) - assert field - self.gguf_writer.add_token_types([field.parts[i].tolist()[0] for i in field.data][:vocab_size]) - - field = neox_reader.get_field(gguf.Keys.Tokenizer.MERGES) - assert field - self.gguf_writer.add_token_merges([bytes(field.parts[i]) for i in field.data]) - - field = neox_reader.get_field(gguf.Keys.Tokenizer.BOS_ID) - self.gguf_writer.add_bos_token_id(field.parts[-1].tolist()[0] if field else 1) - - field = neox_reader.get_field(gguf.Keys.Tokenizer.EOS_ID) - self.gguf_writer.add_eos_token_id(field.parts[-1].tolist()[0] if field else 0) - - field = neox_reader.get_field(gguf.Keys.Tokenizer.UNK_ID) - self.gguf_writer.add_unk_token_id(field.parts[-1].tolist()[0] if field else 0) - - field = neox_reader.get_field(gguf.Keys.Tokenizer.PAD_ID) - self.gguf_writer.add_pad_token_id(field.parts[-1].tolist()[0] if field else 0) + self._set_vocab_builtin("gpt-neox", vocab_size) def set_gguf_parameters(self): d_model = self.find_hparam(["hidden_size", "d_model"]) @@ -2440,7 +2675,6 @@ class MambaModel(Model): # Fail early for models which don't have a block expansion factor of 2 assert d_inner == 2 * d_model - self.gguf_writer.add_name(self.dir_model.name if self.model_name is None else self.model_name) self.gguf_writer.add_context_length(2**20) # arbitrary value; for those who use the default self.gguf_writer.add_embedding_length(d_model) self.gguf_writer.add_feed_forward_length(0) # unused, but seemingly required when loading @@ -2560,7 +2794,7 @@ class JinaBertV2Model(BertModel): yield name, data - def set_vocab(self, *args, **kwargs): + def set_vocab(self): tokenizer_class = 'BertTokenizer' with open(self.dir_model / "tokenizer_config.json", "r", encoding="utf-8") as f: tokenizer_class = json.load(f)['tokenizer_class'] @@ -2576,6 +2810,81 @@ class JinaBertV2Model(BertModel): self.gguf_writer.add_add_eos_token(True) +@Model.register("OpenELMForCausalLM") +class OpenELMModel(Model): + model_arch = gguf.MODEL_ARCH.OPENELM + + @staticmethod + def _make_divisible(v: float | int, divisor: int) -> int: + # ref: https://huggingface.co/apple/OpenELM-270M-Instruct/blob/eb111ff2e6724348e5b905984063d4064d4bc579/configuration_openelm.py#L34-L38 + new_v = max(divisor, int(v + divisor / 2) // divisor * divisor) + # Make sure that round down does not go down by more than 10%. + if new_v < 0.9 * v: + new_v += divisor + return new_v + + def __init__(self, *args, **kwargs): + super().__init__(*args, **kwargs) + + ffn_multipliers: list[float] = self.hparams["ffn_multipliers"] + ffn_dim_divisor: int = self.hparams["ffn_dim_divisor"] + self._n_embd: int = self.hparams["model_dim"] + self._num_kv_heads: list[int] = self.hparams["num_kv_heads"] + self._num_query_heads: list[int] = self.hparams["num_query_heads"] + self._ffn_dims: list[int] = [ + OpenELMModel._make_divisible(multiplier * self._n_embd, ffn_dim_divisor) + for multiplier in ffn_multipliers + ] + assert isinstance(self._num_kv_heads, list) and isinstance(self._num_kv_heads[0], int) + assert isinstance(self._num_query_heads, list) and isinstance(self._num_query_heads[0], int) + + # Uses the tokenizer from meta-llama/Llama-2-7b-hf + def set_vocab(self): + try: + self._set_vocab_sentencepiece() + except FileNotFoundError: + self._set_vocab_builtin("llama-spm", self.hparams["vocab_size"]) + + def set_gguf_parameters(self): + n_embd = self._n_embd + head_dim = self.hparams["head_dim"] + rot_pct = 1.0 + assert self.block_count == len(self._num_kv_heads) + assert self.block_count == len(self._num_query_heads) + assert self.block_count == len(self._ffn_dims) + + self.gguf_writer.add_block_count(self.block_count) + self.gguf_writer.add_context_length(self.hparams["max_context_length"]) + self.gguf_writer.add_embedding_length(n_embd) + self.gguf_writer.add_feed_forward_length(self._ffn_dims) + self.gguf_writer.add_head_count(self._num_query_heads) + self.gguf_writer.add_head_count_kv(self._num_kv_heads) + self.gguf_writer.add_rope_freq_base(self.hparams["rope_freq_constant"]) + # https://huggingface.co/apple/OpenELM-270M-Instruct/blob/c401df2/modeling_openelm.py#L30 + self.gguf_writer.add_layer_norm_rms_eps(1e-6) + self.gguf_writer.add_rope_dimension_count(int(rot_pct * head_dim)) + self.gguf_writer.add_key_length(head_dim) + self.gguf_writer.add_value_length(head_dim) + self.gguf_writer.add_file_type(self.ftype) + + def find_hparam(self, keys: Iterable[str], optional: bool = False) -> Any: + if "n_layers" in keys: + return self.hparams["num_transformer_layers"] + + return super().find_hparam(keys, optional) + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + + # split ff + if bid is not None and name == f"transformer.layers.{bid}.ffn.proj_1.weight": + ff_dim = self._ffn_dims[bid] + yield (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_GATE, bid), data_torch[:ff_dim]) + yield (self.format_tensor_name(gguf.MODEL_TENSOR.FFN_UP, bid), data_torch[ff_dim:]) + return + + yield (self.map_tensor_name(name), data_torch) + + @Model.register("ArcticForCausalLM") class ArcticModel(Model): model_arch = gguf.MODEL_ARCH.ARCTIC @@ -2600,7 +2909,7 @@ class ArcticModel(Model): tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)] scores: list[float] = [-10000.0] * vocab_size - toktypes: list[int] = [SentencePieceTokenTypes.UNKNOWN] * vocab_size + toktypes: list[int] = [SentencePieceTokenTypes.UNUSED] * vocab_size for token_id in range(tokenizer.vocab_size()): @@ -2633,7 +2942,7 @@ class ArcticModel(Model): added_tokens_decoder = tokenizer_config_json["added_tokens_decoder"] for token_id, token_json in added_tokens_decoder.items(): token_id = int(token_id) - if (token_id >= vocab_size): + if token_id >= vocab_size: logger.debug(f'ignore token {token_id}: id is out of range, max={vocab_size - 1}') continue @@ -2717,8 +3026,8 @@ class ArcticModel(Model): return [(self.map_tensor_name(name), data_torch)] - def write_tensors(self): - super().write_tensors() + def prepare_tensors(self): + super().prepare_tensors() if self._experts is not None: # flatten `list[dict[str, Tensor]]` into `list[str]` @@ -2796,8 +3105,8 @@ class DeepseekV2Model(Model): return [(self.map_tensor_name(name), data_torch)] - def write_tensors(self): - super().write_tensors() + def prepare_tensors(self): + super().prepare_tensors() if self._experts is not None: # flatten `list[dict[str, Tensor]]` into `list[str]` @@ -2806,11 +3115,17 @@ class DeepseekV2Model(Model): raise ValueError(f"Unprocessed experts: {experts}") -@Model.register("T5ForConditionalGeneration") @Model.register("T5WithLMHeadModel") +@Model.register("T5ForConditionalGeneration") +@Model.register("MT5ForConditionalGeneration") +@Model.register("UMT5ForConditionalGeneration") class T5Model(Model): model_arch = gguf.MODEL_ARCH.T5 + def __init__(self, *args, **kwargs): + super().__init__(*args, **kwargs) + self.shared_token_embeddings_found = False + def set_vocab(self): # to avoid TypeError: Descriptors cannot be created directly # exception when importing sentencepiece_model_pb2 @@ -2818,17 +3133,29 @@ class T5Model(Model): from sentencepiece import SentencePieceProcessor from sentencepiece import sentencepiece_model_pb2 as model - tokenizer_path = self.dir_model / 'spiece.model' + tokenizer_path = self.dir_model / 'tokenizer.model' + + # many older models use spiece.model tokenizer model filename + if not tokenizer_path.is_file(): + tokenizer_path = self.dir_model / 'spiece.model' if not tokenizer_path.is_file(): raise FileNotFoundError(f"File not found: {tokenizer_path}") - sentencepiece_model = model.ModelProto() + sentencepiece_model = model.ModelProto() # pyright: ignore[reportAttributeAccessIssue] sentencepiece_model.ParseFromString(open(tokenizer_path, "rb").read()) + + # some models like Pile-T5 family use BPE tokenizer instead of Unigram + if sentencepiece_model.trainer_spec.model_type == 2: # BPE + # assure the tokenizer model file name is correct + assert tokenizer_path.name == 'tokenizer.model' + return self._set_vocab_sentencepiece() + else: + assert sentencepiece_model.trainer_spec.model_type == 1 # UNIGRAM + add_prefix = sentencepiece_model.normalizer_spec.add_dummy_prefix remove_whitespaces = sentencepiece_model.normalizer_spec.remove_extra_whitespaces precompiled_charsmap = sentencepiece_model.normalizer_spec.precompiled_charsmap - assert sentencepiece_model.trainer_spec.model_type == 1 # UNIGRAM tokenizer = SentencePieceProcessor() tokenizer.LoadFromFile(str(tokenizer_path)) @@ -2837,7 +3164,7 @@ class T5Model(Model): tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)] scores: list[float] = [-10000.0] * vocab_size - toktypes: list[int] = [SentencePieceTokenTypes.UNKNOWN] * vocab_size + toktypes: list[int] = [SentencePieceTokenTypes.UNUSED] * vocab_size for token_id in range(tokenizer.vocab_size()): piece = tokenizer.IdToPiece(token_id) @@ -2864,7 +3191,7 @@ class T5Model(Model): added_tokens_json = json.load(f) for key in added_tokens_json: token_id = added_tokens_json[key] - if (token_id >= vocab_size): + if token_id >= vocab_size: logger.warning(f'ignore token {token_id}: id is out of range, max={vocab_size - 1}') continue @@ -2897,8 +3224,10 @@ class T5Model(Model): self.gguf_writer.add_add_eos_token(True) def set_gguf_parameters(self): - self.gguf_writer.add_name("T5") - self.gguf_writer.add_context_length(self.hparams["n_positions"]) + if (n_ctx := self.find_hparam(["n_positions"], optional=True)) is None: + logger.warning("Couldn't find context length in config.json, assuming default value of 512") + n_ctx = 512 + self.gguf_writer.add_context_length(n_ctx) self.gguf_writer.add_embedding_length(self.hparams["d_model"]) self.gguf_writer.add_feed_forward_length(self.hparams["d_ff"]) self.gguf_writer.add_block_count(self.hparams["num_layers"]) @@ -2914,16 +3243,292 @@ class T5Model(Model): def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: del bid # unused - # Sometimes T5 and Flan-T5 based models contain "encoder.embed_tokens.weight" tensor or - # "decoder.embed_tokens.weight" tensors that are duplicates of "shared.weight" tensor - # To prevent errors caused by an unnecessary unmapped tensor, skip both of them and use only "shared.weight". - if name == "decoder.embed_tokens.weight" or name == "encoder.embed_tokens.weight": - logger.debug(f"Skipping tensor {name!r} in safetensors so that convert can end normally.") - return [] + # T5 based models contain shared token embeddings tensors saved randomly as either "encoder.embed_tokens.weight", + # "decoder.embed_tokens.weight" or "shared.weight" tensor. In some models there are even multiple of them stored + # in the safetensors files. We use the first tensor from these three as the token embeddings for both encoder + # and decoder and ignore the remaining ones. + if name in ["decoder.embed_tokens.weight", "encoder.embed_tokens.weight", "shared.weight"]: + if not self.shared_token_embeddings_found: + name = "shared.weight" + self.shared_token_embeddings_found = True + else: + logger.debug(f"Skipping shared tensor {name!r} in safetensors so that convert can end normally.") + return [] return [(self.map_tensor_name(name), data_torch)] +@Model.register("JAISLMHeadModel") +class JaisModel(Model): + model_arch = gguf.MODEL_ARCH.JAIS + + def __init__(self, *args, **kwargs): + super().__init__(*args, **kwargs) + + # SwigLU activation + assert self.hparams["activation_function"] == "swiglu" + # ALiBi position embedding + assert self.hparams["position_embedding_type"] == "alibi" + + # Embeddings scale + self.embeddings_scale = 1.0 + # note: For some JAIS flavors, output is tied to (same as) wte in original model + self.output_is_wte = False + if 'mup_embeddings_scale' in self.hparams: + self.output_is_wte = True # Hack (?) + self.embeddings_scale = self.hparams['mup_embeddings_scale'] + elif 'embeddings_scale' in self.hparams: + self.embeddings_scale = self.hparams['embeddings_scale'] + else: + assert False + + self.width_scale = 1.0 + if 'mup_output_alpha' in self.hparams: + assert 'mup_width_scale' in self.hparams + self.width_scale = self.hparams['mup_output_alpha'] * self.hparams['mup_width_scale'] + elif 'width_scale' in self.hparams: + self.width_scale = self.hparams['width_scale'] + else: + assert False + + self.max_alibi_bias = 8.0 + + def set_vocab(self): + self._set_vocab_gpt2() + + def set_gguf_parameters(self): + self.gguf_writer.add_block_count(self.hparams["n_layer"]) + self.gguf_writer.add_context_length(self.hparams["n_positions"]) + self.gguf_writer.add_embedding_length(self.hparams["n_embd"]) + self.gguf_writer.add_feed_forward_length(self.hparams["n_inner"]) + self.gguf_writer.add_head_count(self.hparams["n_head"]) + self.gguf_writer.add_layer_norm_eps(self.hparams["layer_norm_epsilon"]) + self.gguf_writer.add_file_type(self.ftype) + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + del bid # unused + + tensors: list[tuple[str, Tensor]] = [] + + # we don't need these + if name.endswith((".attn.bias")): + return tensors + + if name.endswith(("relative_pe.slopes")): + # Calculate max ALiBi bias (this is the inverse of the ALiBi calculation) + # Some other models has max_alibi_bias spelled out explicitly in the hyperparams, + # but Jais's PyTorch model simply precalculates the slope values and places them + # in relative_pes.slopes + n_head_closest_log2 = 2 ** math.floor(math.log2(self.hparams["n_head"])) + first_val = float(data_torch[0].item()) + self.max_alibi_bias = -round(math.log2(first_val) * n_head_closest_log2) + + return tensors + + if name.endswith((".c_attn.weight", ".c_proj.weight", ".c_fc.weight", ".c_fc2.weight")): + data_torch = data_torch.transpose(1, 0) + + new_name = self.map_tensor_name(name) + + if new_name == self.format_tensor_name(gguf.MODEL_TENSOR.TOKEN_EMBD): + tensors.append((new_name, data_torch * self.embeddings_scale)) + if self.output_is_wte: + tensors.append((self.format_tensor_name(gguf.MODEL_TENSOR.OUTPUT), data_torch * self.width_scale)) + elif new_name == self.format_tensor_name(gguf.MODEL_TENSOR.OUTPUT): + assert not self.output_is_wte + tensors.append((new_name, data_torch * self.width_scale)) + else: + tensors.append((new_name, data_torch)) + + return tensors + + def prepare_tensors(self): + super().prepare_tensors() + self.gguf_writer.add_max_alibi_bias(self.max_alibi_bias) + + +@Model.register("ChatGLMModel", "ChatGLMForConditionalGeneration") +class ChatGLMModel(Model): + model_arch = gguf.MODEL_ARCH.CHATGLM + + def set_vocab_chatglm3(self): + dir_model = self.dir_model + hparams = self.hparams + tokens: list[bytes] = [] + toktypes: list[int] = [] + scores: list[float] = [] + + from transformers import AutoTokenizer + tokenizer = AutoTokenizer.from_pretrained(dir_model, trust_remote_code=True) + vocab_size = hparams.get("padded_vocab_size", len(tokenizer.get_vocab())) + assert max(tokenizer.get_vocab().values()) < vocab_size + role_special_tokens = ["<|system|>", "<|user|>", "<|assistant|>", "<|observation|>"] + special_tokens = ["[MASK]", "[gMASK]", "[sMASK]", "sop", "eop"] + role_special_tokens + for token_id in range(vocab_size): + piece = tokenizer._convert_id_to_token(token_id) + if token_id == 0: + piece = "" + elif token_id == 1: + piece = "" + elif token_id == 2: + piece = "" + + text = piece.encode("utf-8") + score = 0.0 + # Referencing the tokenizer Python implementation(https://huggingface.co/THUDM/chatglm3-6b/blob/main/tokenization_chatglm.py), + # it is only valid if it is less than tokenizer.tokenizer.sp_model.vocab_size() + if len(piece) != 0 and token_id < tokenizer.tokenizer.sp_model.vocab_size(): + score = tokenizer.tokenizer.sp_model.get_score(token_id) + + if token_id >= tokenizer.tokenizer.sp_model.vocab_size(): + if piece in special_tokens: + toktype = SentencePieceTokenTypes.CONTROL + elif len(piece) == 0: + text = f"[PAD{token_id}]".encode("utf-8") + toktype = SentencePieceTokenTypes.UNUSED + else: + toktype = SentencePieceTokenTypes.USER_DEFINED + tokens.append(text) + scores.append(score) + toktypes.append(toktype) + continue + + toktype = SentencePieceTokenTypes.NORMAL + if tokenizer.tokenizer.sp_model.is_unknown(token_id): + toktype = SentencePieceTokenTypes.UNKNOWN + elif tokenizer.tokenizer.sp_model.is_control(token_id): + toktype = SentencePieceTokenTypes.CONTROL + elif tokenizer.tokenizer.sp_model.is_unused(token_id): + toktype = SentencePieceTokenTypes.UNUSED + elif tokenizer.tokenizer.sp_model.is_byte(token_id): + toktype = SentencePieceTokenTypes.BYTE + + tokens.append(text) + scores.append(score) + toktypes.append(toktype) + + self.gguf_writer.add_tokenizer_model("llama") + # glm3 needs prefix and suffix formatted as: + # prompt = "[gMASK]sop<|user|>\n" + prompt + "<|assistant|>" + self.gguf_writer.add_tokenizer_pre("chatglm-spm") + self.gguf_writer.add_token_list(tokens) + self.gguf_writer.add_token_scores(scores) + self.gguf_writer.add_token_types(toktypes) + + special_vocab = gguf.SpecialVocab(self.dir_model, n_vocab=len(tokens)) + special_vocab.add_to_gguf(self.gguf_writer) + + @staticmethod + def token_bytes_to_string(b): + from transformers.models.gpt2.tokenization_gpt2 import bytes_to_unicode + byte_encoder = bytes_to_unicode() + return ''.join([byte_encoder[ord(char)] for char in b.decode('latin-1')]) + + @staticmethod + def bpe(mergeable_ranks: dict[bytes, int], token: bytes, max_rank: int | None = None) -> list[bytes]: + parts = [bytes([b]) for b in token] + while True: + min_idx = None + min_rank = None + for i, pair in enumerate(zip(parts[:-1], parts[1:])): + rank = mergeable_ranks.get(pair[0] + pair[1]) + if rank is not None and (min_rank is None or rank < min_rank): + min_idx = i + min_rank = rank + if min_rank is None or (max_rank is not None and min_rank >= max_rank): + break + assert min_idx is not None + parts = parts[:min_idx] + [parts[min_idx] + parts[min_idx + 1]] + parts[min_idx + 2:] + return parts + + def set_vocab(self): + if "THUDM/chatglm3-6b" in self.hparams.get("_name_or_path", ""): + self.set_vocab_chatglm3() + return + + dir_model = self.dir_model + hparams = self.hparams + tokens: list[str] = [] + toktypes: list[int] = [] + + from transformers import AutoTokenizer + tokenizer = AutoTokenizer.from_pretrained(dir_model, trust_remote_code=True) + vocab_size = hparams["padded_vocab_size"] + assert max(tokenizer.get_vocab().values()) < vocab_size + + tokpre = self.get_vocab_base_pre(tokenizer) + + merges = [] + vocab = {} + mergeable_ranks = tokenizer.mergeable_ranks + for token, rank in mergeable_ranks.items(): + vocab[ChatGLMModel.token_bytes_to_string(token)] = rank + if len(token) == 1: + continue + merged = ChatGLMModel.bpe(mergeable_ranks, token, max_rank=rank) + assert len(merged) >= 2 and len(merged) <= 7 + merges.append(' '.join(map(ChatGLMModel.token_bytes_to_string, merged))) + + # for this kind of tokenizer, added_vocab is not a subset of vocab, so they need to be combined + added_vocab = tokenizer.get_added_vocab() + reverse_vocab = {id_ : encoded_tok for encoded_tok, id_ in {**vocab, **added_vocab}.items()} + + for i in range(vocab_size): + if i not in reverse_vocab: + tokens.append(f"[PAD{i}]") + toktypes.append(gguf.TokenType.UNUSED) + elif reverse_vocab[i] in added_vocab: + tokens.append(reverse_vocab[i]) + if tokenizer.added_tokens_decoder[i].special: + toktypes.append(gguf.TokenType.CONTROL) + else: + toktypes.append(gguf.TokenType.USER_DEFINED) + else: + tokens.append(reverse_vocab[i]) + toktypes.append(gguf.TokenType.NORMAL) + + self.gguf_writer.add_tokenizer_model("gpt2") + self.gguf_writer.add_tokenizer_pre(tokpre) + self.gguf_writer.add_token_list(tokens) + self.gguf_writer.add_token_types(toktypes) + + special_vocab = gguf.SpecialVocab(dir_model, load_merges=False) + special_vocab.merges = merges + # only add special tokens when they were not already loaded from config.json + special_vocab._set_special_token("eos", tokenizer.get_added_vocab()["<|endoftext|>"]) + special_vocab._set_special_token("eot", tokenizer.get_added_vocab()["<|user|>"]) + # this one is usually not in config.json anyway + special_vocab._set_special_token("unk", tokenizer.get_added_vocab()["<|endoftext|>"]) + special_vocab.add_to_gguf(self.gguf_writer) + + def set_gguf_parameters(self): + n_embed = self.hparams.get("hidden_size", self.hparams.get("n_embed")) + n_head = self.hparams.get("n_head", self.hparams.get("num_attention_heads")) + n_head_kv = self.hparams.get("multi_query_group_num", n_head) + self.gguf_writer.add_context_length(self.hparams.get("seq_length", n_embed)) + self.gguf_writer.add_embedding_length(n_embed) + self.gguf_writer.add_feed_forward_length(self.hparams.get("ffn_hidden_size", 4 * n_embed)) + self.gguf_writer.add_block_count(self.hparams["num_layers"]) + self.gguf_writer.add_head_count(n_head) + self.gguf_writer.add_head_count_kv(n_head_kv) + self.gguf_writer.add_layer_norm_rms_eps(self.hparams["layernorm_epsilon"]) + self.gguf_writer.add_file_type(self.ftype) + self.gguf_writer.add_rope_dimension_count(64) + self.gguf_writer.add_add_bos_token(False) + rope_freq = 10000 + if "rope_ratio" in self.hparams: + rope_freq = rope_freq * self.hparams["rope_ratio"] + self.gguf_writer.add_rope_freq_base(rope_freq) + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + del bid # unused + + if name.endswith(".rotary_pos_emb.inv_freq"): + return [] + + name = name.removeprefix("transformer.") + return [(self.map_tensor_name(name), data_torch)] + ###### CONVERSION LOGIC ###### @@ -2940,19 +3545,46 @@ class LazyTorchTensor(gguf.LazyBase): torch.float32: np.float32, } + # used for safetensors slices + # ref: https://github.com/huggingface/safetensors/blob/079781fd0dc455ba0fe851e2b4507c33d0c0d407/bindings/python/src/lib.rs#L1046 + # TODO: uncomment U64, U32, and U16, ref: https://github.com/pytorch/pytorch/issues/58734 + _dtype_str_map: dict[str, torch.dtype] = { + "F64": torch.float64, + "F32": torch.float32, + "BF16": torch.bfloat16, + "F16": torch.float16, + # "U64": torch.uint64, + "I64": torch.int64, + # "U32": torch.uint32, + "I32": torch.int32, + # "U16": torch.uint16, + "I16": torch.int16, + "U8": torch.uint8, + "I8": torch.int8, + "BOOL": torch.bool, + "F8_E4M3": torch.float8_e4m3fn, + "F8_E5M2": torch.float8_e5m2, + } + def numpy(self) -> gguf.LazyNumpyTensor: dtype = self._dtype_map[self.dtype] return gguf.LazyNumpyTensor( meta=gguf.LazyNumpyTensor.meta_with_dtype_and_shape(dtype, self.shape), - lazy=self._lazy, args=(self,), - func=(lambda s: s[0].numpy()) + func=(lambda s: s.numpy()) ) @classmethod - def meta_with_dtype_and_shape(cls, dtype: torch.dtype, shape: torch.Size) -> Tensor: + def meta_with_dtype_and_shape(cls, dtype: torch.dtype, shape: tuple[int, ...]) -> Tensor: return torch.empty(size=shape, dtype=dtype, device="meta") + @classmethod + def from_safetensors_slice(cls, st_slice: Any) -> Tensor: + dtype = cls._dtype_str_map[st_slice.get_dtype()] + shape: tuple[int, ...] = tuple(st_slice.get_shape()) + lazy = cls(meta=cls.meta_with_dtype_and_shape(dtype, shape), args=(st_slice,), func=lambda s: s[:]) + return cast(torch.Tensor, lazy) + @classmethod def __torch_function__(cls, func, types, args=(), kwargs=None): del types # unused @@ -2963,7 +3595,7 @@ class LazyTorchTensor(gguf.LazyBase): if func is torch.Tensor.numpy: return args[0].numpy() - return LazyTorchTensor._wrap_fn(func)(*args, **kwargs) + return cls._wrap_fn(func)(*args, **kwargs) def parse_args() -> argparse.Namespace: @@ -2973,10 +3605,6 @@ def parse_args() -> argparse.Namespace: "--vocab-only", action="store_true", help="extract only the vocab", ) - parser.add_argument( - "--awq-path", type=Path, default=None, - help="Path to scale awq cache file", - ) parser.add_argument( "--outfile", type=Path, help="path to write to; default: based on input. {ftype} will be replaced by the outtype.", @@ -3025,6 +3653,10 @@ def parse_args() -> argparse.Namespace: "--no-tensor-first-split", action="store_true", help="do not add tensors to the first split (disabled by default)" ) + parser.add_argument( + "--metadata", type=Path, + help="Specify the path for an authorship metadata override file" + ) return parser.parse_args() @@ -3050,23 +3682,13 @@ def split_str_to_n_bytes(split_str: str) -> int: def main() -> None: args = parse_args() - logging.basicConfig(level=logging.DEBUG if args.verbose else logging.INFO) + if args.verbose: + logging.basicConfig(level=logging.DEBUG) + else: + logging.basicConfig(level=logging.INFO) dir_model = args.model - if args.awq_path: - sys.path.insert(1, str(Path(__file__).parent / 'awq-py')) - from awq.apply_awq import add_scale_weights # type: ignore[import-not-found] - tmp_model_path = args.model / "weighted_model" - dir_model = tmp_model_path - if tmp_model_path.is_dir(): - logger.info(f"{tmp_model_path} exists as a weighted model.") - else: - tmp_model_path.mkdir(parents=True, exist_ok=True) - logger.info("Saving new weighted model ...") - add_scale_weights(str(args.model), str(args.awq_path), str(tmp_model_path)) - logger.info(f"Saved weighted model at {tmp_model_path}.") - if not dir_model.is_dir(): logger.error(f'Error: {args.model} is not a directory') sys.exit(1) @@ -3079,48 +3701,47 @@ def main() -> None: "auto": gguf.LlamaFileType.GUESSED, } - if args.use_temp_file and (args.split_max_tensors > 0 or args.split_max_size != "0"): + is_split = args.split_max_tensors > 0 or args.split_max_size != "0" + if args.use_temp_file and is_split: logger.error("Error: Cannot use temp file when splitting") sys.exit(1) if args.outfile is not None: fname_out = args.outfile else: - # output in the same directory as the model by default - fname_out = dir_model / 'ggml-model-{ftype}.gguf' + fname_out = dir_model logger.info(f"Loading model: {dir_model.name}") hparams = Model.load_hparams(dir_model) with torch.inference_mode(): + output_type = ftype_map[args.outtype] + model_architecture = hparams["architectures"][0] + try: - model_class = Model.from_model_architecture(hparams["architectures"][0]) + model_class = Model.from_model_architecture(model_architecture) except NotImplementedError: - logger.error(f"Model {hparams['architectures'][0]} is not supported") + logger.error(f"Model {model_architecture} is not supported") sys.exit(1) - model_instance = model_class(dir_model, ftype_map[args.outtype], fname_out, args.bigendian, args.use_temp_file, - args.no_lazy, args.model_name, split_max_tensors=args.split_max_tensors, + model_instance = model_class(dir_model=dir_model, ftype=output_type, fname_out=fname_out, + is_big_endian=args.bigendian, use_temp_file=args.use_temp_file, + eager=args.no_lazy, + metadata_override=args.metadata, model_name=args.model_name, + split_max_tensors=args.split_max_tensors, split_max_size=split_str_to_n_bytes(args.split_max_size), dry_run=args.dry_run, small_first_shard=args.no_tensor_first_split) - logger.info("Set model parameters") - model_instance.set_gguf_parameters() - - logger.info("Set model tokenizer") - model_instance.set_vocab() - - model_instance.gguf_writer.add_quantization_version(gguf.GGML_QUANT_VERSION) - if args.vocab_only: logger.info("Exporting model vocab...") model_instance.write_vocab() - logger.info("Model vocab successfully exported.") + logger.info(f"Model vocab successfully exported to {model_instance.fname_out}") else: logger.info("Exporting model...") model_instance.write() - logger.info("Model successfully exported.") + out_path = f"{model_instance.fname_out.parent}{os.sep}" if is_split else model_instance.fname_out + logger.info(f"Model successfully exported to {out_path}") if __name__ == '__main__': diff --git a/convert-hf-to-gguf-update.py b/convert_hf_to_gguf_update.py similarity index 79% rename from convert-hf-to-gguf-update.py rename to convert_hf_to_gguf_update.py index 67598b561..d5a2d925e 100755 --- a/convert-hf-to-gguf-update.py +++ b/convert_hf_to_gguf_update.py @@ -2,7 +2,7 @@ # -*- coding: utf-8 -*- # This script downloads the tokenizer models of the specified models from Huggingface and -# generates the get_vocab_base_pre() function for convert-hf-to-gguf.py +# generates the get_vocab_base_pre() function for convert_hf_to_gguf.py # # This is necessary in order to analyze the type of pre-tokenizer used by the model and # provide the necessary information to llama.cpp via the GGUF header in order to implement @@ -15,9 +15,9 @@ # - Add a new model to the "models" list # - Run the script with your huggingface token: # -# python3 convert-hf-to-gguf-update.py +# python3 convert_hf_to_gguf_update.py # -# - Copy-paste the generated get_vocab_base_pre() function into convert-hf-to-gguf.py +# - Copy-paste the generated get_vocab_base_pre() function into convert_hf_to_gguf.py # - Update llama.cpp with the new pre-tokenizer if necessary # # TODO: generate tokenizer tests for llama.cpp @@ -37,7 +37,7 @@ from enum import IntEnum, auto from transformers import AutoTokenizer logging.basicConfig(level=logging.DEBUG) -logger = logging.getLogger("convert-hf-to-gguf-update") +logger = logging.getLogger("convert_hf_to_gguf_update") sess = requests.Session() @@ -45,20 +45,21 @@ class TOKENIZER_TYPE(IntEnum): SPM = auto() BPE = auto() WPM = auto() + UGM = auto() # TODO: this string has to exercise as much pre-tokenizer functionality as possible # will be updated with time - contributions welcome -chktxt = '\n \n\n \n\n\n \t \t\t \t\n \n \n \n \n🚀 (normal) 😶‍🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български \'\'\'\'\'\'```````\"\"\"\"......!!!!!!?????? I\'ve been \'told he\'s there, \'RE you sure? \'M not sure I\'ll make it, \'D you like some tea? We\'Ve a\'lL' +CHK_TXT = '\n \n\n \n\n\n \t \t\t \t\n \n \n \n \n🚀 (normal) 😶‍🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български \'\'\'\'\'\'```````\"\"\"\"......!!!!!!?????? I\'ve been \'told he\'s there, \'RE you sure? \'M not sure I\'ll make it, \'D you like some tea? We\'Ve a\'lL' if len(sys.argv) == 2: token = sys.argv[1] if not token.startswith("hf_"): logger.info("Huggingface token seems invalid") - logger.info("Usage: python convert-hf-to-gguf-update.py ") + logger.info("Usage: python convert_hf_to_gguf_update.py ") sys.exit(1) else: - logger.info("Usage: python convert-hf-to-gguf-update.py ") + logger.info("Usage: python convert_hf_to_gguf_update.py ") sys.exit(1) # TODO: add models here, base models preferred @@ -85,6 +86,14 @@ models = [ {"name": "smaug-bpe", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/abacusai/Smaug-Llama-3-70B-Instruct", }, {"name": "poro-chat", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/LumiOpen/Poro-34B-chat", }, {"name": "jina-v2-code", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-code", }, + {"name": "viking", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/LumiOpen/Viking-7B", }, # Also used for Viking 13B and 33B + {"name": "gemma", "tokt": TOKENIZER_TYPE.SPM, "repo": "https://huggingface.co/google/gemma-2b", }, + {"name": "gemma-2", "tokt": TOKENIZER_TYPE.SPM, "repo": "https://huggingface.co/google/gemma-2-9b", }, + {"name": "jais", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/core42/jais-13b", }, + {"name": "t5", "tokt": TOKENIZER_TYPE.UGM, "repo": "https://huggingface.co/google-t5/t5-small", }, + {"name": "codeshell", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/WisdomShell/CodeShell-7B", }, + {"name": "tekken", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/mistralai/Mistral-Nemo-Base-2407", }, + {"name": "smollm", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/HuggingFaceTB/SmolLM-135M", }, ] @@ -93,8 +102,8 @@ def download_file_with_auth(url, token, save_path): response = sess.get(url, headers=headers) response.raise_for_status() os.makedirs(os.path.dirname(save_path), exist_ok=True) - with open(save_path, 'wb') as f: - f.write(response.content) + with open(save_path, 'wb') as downloaded_file: + downloaded_file.write(response.content) logger.info(f"File {save_path} downloaded successfully") @@ -106,9 +115,13 @@ def download_model(model): os.makedirs(f"models/tokenizers/{name}", exist_ok=True) files = ["config.json", "tokenizer.json", "tokenizer_config.json"] + if tokt == TOKENIZER_TYPE.SPM: files.append("tokenizer.model") + if tokt == TOKENIZER_TYPE.UGM: + files.append("spiece.model") + for file in files: save_path = f"models/tokenizers/{name}/{file}" if os.path.isfile(save_path): @@ -124,14 +137,14 @@ for model in models: logger.error(f"Failed to download model {model['name']}. Error: {e}") -# generate the source code for the convert-hf-to-gguf.py:get_vocab_base_pre() function: +# generate the source code for the convert_hf_to_gguf.py:get_vocab_base_pre() function: src_ifs = "" for model in models: name = model["name"] tokt = model["tokt"] - if tokt == TOKENIZER_TYPE.SPM: + if tokt == TOKENIZER_TYPE.SPM or tokt == TOKENIZER_TYPE.UGM: continue # Skip if the tokenizer folder does not exist or there are other download issues previously @@ -141,12 +154,15 @@ for model in models: # create the tokenizer try: - tokenizer = AutoTokenizer.from_pretrained(f"models/tokenizers/{name}") + if name == "t5": + tokenizer = AutoTokenizer.from_pretrained(f"models/tokenizers/{name}", use_fast=False) + else: + tokenizer = AutoTokenizer.from_pretrained(f"models/tokenizers/{name}") except OSError as e: logger.error(f"Error loading tokenizer for model {name}. The model may not exist or is not accessible with the provided token. Error: {e}") continue # Skip to the next model if the tokenizer can't be loaded - chktok = tokenizer.encode(chktxt) + chktok = tokenizer.encode(CHK_TXT) chkhsh = sha256(str(chktok).encode()).hexdigest() logger.info(f"model: {name}") @@ -178,7 +194,7 @@ src_func = f""" # we will use this unique identifier to write a "tokenizer.ggml.pre" entry in the GGUF file which we can # use in llama.cpp to implement the same pre-tokenizer - chktxt = {repr(chktxt)} + chktxt = {repr(CHK_TXT)} chktok = tokenizer.encode(chktxt) chkhsh = sha256(str(chktok).encode()).hexdigest() @@ -188,7 +204,7 @@ src_func = f""" res = None - # NOTE: if you get an error here, you need to update the convert-hf-to-gguf-update.py script + # NOTE: if you get an error here, you need to update the convert_hf_to_gguf_update.py script # or pull the latest version of the model from Huggingface # don't edit the hashes manually! {src_ifs} @@ -197,9 +213,9 @@ src_func = f""" logger.warning("**************************************************************************************") logger.warning("** WARNING: The BPE pre-tokenizer was not recognized!") logger.warning("** There are 2 possible reasons for this:") - logger.warning("** - the model has not been added to convert-hf-to-gguf-update.py yet") + logger.warning("** - the model has not been added to convert_hf_to_gguf_update.py yet") logger.warning("** - the pre-tokenization config has changed upstream") - logger.warning("** Check your model files and convert-hf-to-gguf-update.py and update them accordingly.") + logger.warning("** Check your model files and convert_hf_to_gguf_update.py and update them accordingly.") logger.warning("** ref: https://github.com/ggerganov/llama.cpp/pull/6920") logger.warning("**") logger.warning(f"** chkhsh: {{chkhsh}}") @@ -213,7 +229,7 @@ src_func = f""" return res """ -convert_py_pth = pathlib.Path("convert-hf-to-gguf.py") +convert_py_pth = pathlib.Path("convert_hf_to_gguf.py") convert_py = convert_py_pth.read_text(encoding="utf-8") convert_py = re.sub( r"(# Marker: Start get_vocab_base_pre)(.+?)( +# Marker: End get_vocab_base_pre)", @@ -224,7 +240,7 @@ convert_py = re.sub( convert_py_pth.write_text(convert_py, encoding="utf-8") -logger.info("+++ convert-hf-to-gguf.py was updated") +logger.info("+++ convert_hf_to_gguf.py was updated") # generate tests for each tokenizer model @@ -262,6 +278,7 @@ tests = [ "\n =", "' era", "Hello, y'all! How are you 😁 ?我想在apple工作1314151天～", + "!!!!!!", "3", "33", "333", @@ -271,8 +288,9 @@ tests = [ "3333333", "33333333", "333333333", - # "Cửa Việt", # llama-bpe fails on this - chktxt, + "Cửa Việt", # llama-bpe fails on this + " discards", + CHK_TXT, ] # write the tests to ./models/ggml-vocab-{name}.gguf.inp @@ -299,7 +317,10 @@ for model in models: # create the tokenizer try: - tokenizer = AutoTokenizer.from_pretrained(f"models/tokenizers/{name}") + if name == "t5": + tokenizer = AutoTokenizer.from_pretrained(f"models/tokenizers/{name}", use_fast=False) + else: + tokenizer = AutoTokenizer.from_pretrained(f"models/tokenizers/{name}") except OSError as e: logger.error(f"Failed to load tokenizer for model {name}. Error: {e}") continue # Skip this model and continue with the next one in the loop @@ -325,6 +346,6 @@ logger.info("\nRun the following commands to generate the vocab files for testin for model in models: name = model["name"] - print(f"python3 convert-hf-to-gguf.py models/tokenizers/{name}/ --outfile models/ggml-vocab-{name}.gguf --vocab-only") # noqa: NP100 + print(f"python3 convert_hf_to_gguf.py models/tokenizers/{name}/ --outfile models/ggml-vocab-{name}.gguf --vocab-only") # noqa: NP100 logger.info("\n") diff --git a/convert-llama-ggml-to-gguf.py b/convert_llama_ggml_to_gguf.py similarity index 99% rename from convert-llama-ggml-to-gguf.py rename to convert_llama_ggml_to_gguf.py index 9349de3b3..7b00b4398 100755 --- a/convert-llama-ggml-to-gguf.py +++ b/convert_llama_ggml_to_gguf.py @@ -132,6 +132,10 @@ class Tensor: class GGMLModel: + + file_format: GGMLFormat + format_version: int + def __init__(self): self.hyperparameters = None self.vocab = None @@ -290,7 +294,7 @@ class GGMLToGGUF: if self.vocab_override is not None: vo = self.vocab_override logger.info('* Adding vocab item(s)') - for (idx, (vbytes, score, ttype)) in enumerate(vo.all_tokens()): + for (_, (vbytes, score, ttype)) in enumerate(vo.all_tokens()): tokens.append(vbytes) scores.append(score) toktypes.append(ttype) @@ -354,7 +358,8 @@ class GGMLToGGUF: def handle_metadata(cfg, hp): - import convert + import examples.convert_legacy_llama as convert + assert cfg.model_metadata_dir.is_dir(), 'Metadata dir is not a directory' hf_config_path = cfg.model_metadata_dir / "config.json" orig_config_path = cfg.model_metadata_dir / "params.json" diff --git a/convert_lora_to_gguf.py b/convert_lora_to_gguf.py new file mode 100755 index 000000000..a88d0d4a9 --- /dev/null +++ b/convert_lora_to_gguf.py @@ -0,0 +1,393 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8 -*- + +from __future__ import annotations + +from dataclasses import dataclass +import logging +import argparse +import os +import sys +import json +from math import prod +from pathlib import Path +from typing import TYPE_CHECKING, Any, Callable, Iterable, Iterator, Sequence, SupportsIndex, cast + +import torch + +if TYPE_CHECKING: + from torch import Tensor + +if 'NO_LOCAL_GGUF' not in os.environ: + sys.path.insert(1, str(Path(__file__).parent / 'gguf-py')) +import gguf + +# reuse model definitions from convert_hf_to_gguf.py +from convert_hf_to_gguf import LazyTorchTensor, Model + +logger = logging.getLogger("lora-to-gguf") + + +@dataclass +class PartialLoraTensor: + A: Tensor | None = None + B: Tensor | None = None + + +# magic to support tensor shape modifications and splitting +class LoraTorchTensor: + _lora_A: Tensor # (n_rank, row_size) + _lora_B: Tensor # (col_size, n_rank) + _rank: int + + def __init__(self, A: Tensor, B: Tensor): + assert len(A.shape) == len(B.shape) + assert A.shape[-2] == B.shape[-1] + if A.dtype != B.dtype: + A = A.to(torch.float32) + B = B.to(torch.float32) + self._lora_A = A + self._lora_B = B + self._rank = B.shape[-1] + + def get_lora_A_B(self) -> tuple[Tensor, Tensor]: + return (self._lora_A, self._lora_B) + + def __getitem__( + self, + indices: ( + SupportsIndex + | slice + | tuple[SupportsIndex | slice | Tensor, ...] # TODO: add ellipsis in the type signature + ), + ) -> LoraTorchTensor: + shape = self.shape + if isinstance(indices, SupportsIndex): + if len(shape) > 2: + return LoraTorchTensor(self._lora_A[indices], self._lora_B[indices]) + else: + raise NotImplementedError # can't return a vector + elif isinstance(indices, slice): + if len(shape) > 2: + return LoraTorchTensor(self._lora_A[indices], self._lora_B[indices]) + else: + return LoraTorchTensor(self._lora_A, self._lora_B[indices]) + elif isinstance(indices, tuple): + assert len(indices) > 0 + if indices[-1] is Ellipsis: + return self[indices[:-1]] + # expand ellipsis + indices = tuple( + u + for v in ( + ( + (slice(None, None) for _ in range(len(indices) - 1)) + if i is Ellipsis + else (i,) + ) + for i in indices + ) + for u in v + ) + + if len(indices) < len(shape): + indices = (*indices, *(slice(None, None) for _ in range(len(indices), len(shape)))) + + # TODO: make sure this is correct + indices_A = ( + *( + ( + j.__index__() % self._lora_A.shape[i] + if isinstance(j, SupportsIndex) + else slice(None, None) + ) + for i, j in enumerate(indices[:-2]) + ), + slice(None, None), + indices[-1], + ) + indices_B = indices[:-1] + return LoraTorchTensor(self._lora_A[indices_A], self._lora_B[indices_B]) + else: + raise NotImplementedError # unknown indice type + + @property + def dtype(self) -> torch.dtype: + assert self._lora_A.dtype == self._lora_B.dtype + return self._lora_A.dtype + + @property + def shape(self) -> tuple[int, ...]: + assert len(self._lora_A.shape) == len(self._lora_B.shape) + return (*self._lora_B.shape[:-1], self._lora_A.shape[-1]) + + def size(self, dim=None): + assert dim is None + return self.shape + + def reshape(self, *shape: int | tuple[int, ...]) -> LoraTorchTensor: + if isinstance(shape[0], tuple): + new_shape: tuple[int, ...] = shape[0] + else: + new_shape = cast(tuple[int, ...], shape) + orig_shape = self.shape + if len(new_shape) < 2: + raise NotImplementedError # can't become a vector + + # expand -1 in the shape + if any(dim == -1 for dim in new_shape): + n_elems = prod(orig_shape) + n_new_elems = prod(dim if dim != -1 else 1 for dim in new_shape) + assert n_elems % n_new_elems == 0 + new_shape = (*(dim if dim != -1 else n_elems // n_new_elems for dim in new_shape),) + + if new_shape[-1] != orig_shape[-1]: + raise NotImplementedError # can't reshape the row size trivially + + shape_A = (*(1 for _ in new_shape[:-2]), self._rank, orig_shape[-1]) + shape_B = (*new_shape[:-1], self._rank) + return LoraTorchTensor( + self._lora_A.reshape(shape_A), + self._lora_B.reshape(shape_B), + ) + + def reshape_as(self, other: Tensor) -> LoraTorchTensor: + return self.reshape(*other.shape) + + def view(self, *size: int) -> LoraTorchTensor: + return self.reshape(*size) + + def permute(self, *dims: int) -> LoraTorchTensor: + shape = self.shape + dims = tuple(dim - len(shape) if dim >= 0 else dim for dim in dims) + if dims[-1] == -1: + # TODO: support higher dimensional A shapes bigger than 1 + assert all(dim == 1 for dim in self._lora_A.shape[:-2]) + return LoraTorchTensor(self._lora_A, self._lora_B.permute(*dims)) + if len(shape) == 2 and dims[-1] == -2 and dims[-2] == -1: + return LoraTorchTensor(self._lora_B.permute(*dims), self._lora_A.permute(*dims)) + else: + # TODO: compose the above two + raise NotImplementedError + + def transpose(self, dim0: int, dim1: int) -> LoraTorchTensor: + shape = self.shape + dims = [i for i in range(len(shape))] + dims[dim0], dims[dim1] = dims[dim1], dims[dim0] + return self.permute(*dims) + + def swapaxes(self, axis0: int, axis1: int) -> LoraTorchTensor: + return self.transpose(axis0, axis1) + + def to(self, *args, **kwargs): + return LoraTorchTensor(self._lora_A.to(*args, **kwargs), self._lora_B.to(*args, **kwargs)) + + @classmethod + def __torch_function__(cls, func: Callable, types, args=(), kwargs=None): + del types # unused + + if kwargs is None: + kwargs = {} + + if func is torch.permute: + return type(args[0]).permute(*args, **kwargs) + elif func is torch.reshape: + return type(args[0]).reshape(*args, **kwargs) + elif func is torch.stack: + assert isinstance(args[0], Sequence) + dim = kwargs.get("dim", 0) + assert dim == 0 + return LoraTorchTensor( + torch.stack([a._lora_A for a in args[0]], dim), + torch.stack([b._lora_B for b in args[0]], dim), + ) + elif func is torch.cat: + assert isinstance(args[0], Sequence) + dim = kwargs.get("dim", 0) + assert dim == 0 + if len(args[0][0].shape) > 2: + return LoraTorchTensor( + torch.cat([a._lora_A for a in args[0]], dim), + torch.cat([b._lora_B for b in args[0]], dim), + ) + elif all(torch.equal(args[0][0]._lora_A, t._lora_A) for t in args[0][1:]): + return LoraTorchTensor( + args[0][0]._lora_A, + torch.cat([b._lora_B for b in args[0]], dim), + ) + else: + raise NotImplementedError + else: + raise NotImplementedError + + +def get_base_tensor_name(lora_tensor_name: str) -> str: + base_name = lora_tensor_name.replace("base_model.model.", "") + base_name = base_name.replace(".lora_A.weight", ".weight") + base_name = base_name.replace(".lora_B.weight", ".weight") + return base_name + + +def parse_args() -> argparse.Namespace: + parser = argparse.ArgumentParser( + description="Convert a huggingface PEFT LoRA adapter to a GGML compatible file") + parser.add_argument( + "--outfile", type=Path, + help="path to write to; default: based on input. {ftype} will be replaced by the outtype.", + ) + parser.add_argument( + "--outtype", type=str, choices=["f32", "f16", "bf16", "q8_0", "auto"], default="f16", + help="output format - use f32 for float32, f16 for float16, bf16 for bfloat16, q8_0 for Q8_0, auto for the highest-fidelity 16-bit float type depending on the first loaded tensor type", + ) + parser.add_argument( + "--bigendian", action="store_true", + help="model is executed on big endian machine", + ) + parser.add_argument( + "--no-lazy", action="store_true", + help="use more RAM by computing all outputs before writing (use in case lazy evaluation is broken)", + ) + parser.add_argument( + "--verbose", action="store_true", + help="increase output verbosity", + ) + parser.add_argument( + "--dry-run", action="store_true", + help="only print out what will be done, without writing any new files", + ) + parser.add_argument( + "--base", type=Path, required=True, + help="directory containing base model file", + ) + parser.add_argument( + "lora_path", type=Path, + help="directory containing LoRA adapter file", + ) + + return parser.parse_args() + + +if __name__ == '__main__': + args = parse_args() + logging.basicConfig(level=logging.DEBUG if args.verbose else logging.INFO) + + ftype_map: dict[str, gguf.LlamaFileType] = { + "f32": gguf.LlamaFileType.ALL_F32, + "f16": gguf.LlamaFileType.MOSTLY_F16, + "bf16": gguf.LlamaFileType.MOSTLY_BF16, + "q8_0": gguf.LlamaFileType.MOSTLY_Q8_0, + "auto": gguf.LlamaFileType.GUESSED, + } + + ftype = ftype_map[args.outtype] + + dir_base_model: Path = args.base + dir_lora: Path = args.lora_path + lora_config = dir_lora / "adapter_config.json" + input_model = dir_lora / "adapter_model.safetensors" + + if args.outfile is not None: + fname_out = args.outfile + else: + # output in the same directory as the model by default + fname_out = dir_lora + + if os.path.exists(input_model): + # lazy import load_file only if lora is in safetensors format. + from safetensors.torch import load_file + + lora_model = load_file(input_model, device="cpu") + else: + input_model = os.path.join(dir_lora, "adapter_model.bin") + lora_model = torch.load(input_model, map_location="cpu", weights_only=True) + + # load base model + logger.info(f"Loading base model: {dir_base_model.name}") + hparams = Model.load_hparams(dir_base_model) + with torch.inference_mode(): + try: + model_class = Model.from_model_architecture(hparams["architectures"][0]) + except NotImplementedError: + logger.error(f"Model {hparams['architectures'][0]} is not supported") + sys.exit(1) + + class LoraModel(model_class): + model_arch = model_class.model_arch + + lora_alpha: float + + def __init__(self, *args, dir_lora_model: Path, lora_alpha: float, **kwargs): + + super().__init__(*args, **kwargs) + + self.dir_model_card = dir_lora_model + self.lora_alpha = float(lora_alpha) + + def set_type(self): + self.gguf_writer.add_type(gguf.GGUFType.ADAPTER) + self.gguf_writer.add_string(gguf.Keys.Adapter.TYPE, "lora") + + def set_gguf_parameters(self): + self.gguf_writer.add_float32(gguf.Keys.Adapter.LORA_ALPHA, self.lora_alpha) + super().set_gguf_parameters() + + def get_tensors(self) -> Iterator[tuple[str, Tensor]]: + tensor_map: dict[str, PartialLoraTensor] = {} + + for name, tensor in lora_model.items(): + if self.lazy: + tensor = LazyTorchTensor.from_eager(tensor) + base_name = get_base_tensor_name(name) + is_lora_a = ".lora_A.weight" in name + is_lora_b = ".lora_B.weight" in name + if not is_lora_a and not is_lora_b: + if ".base_layer.weight" in name: + continue + logger.error(f"Unexpected name '{name}': Not a lora_A or lora_B tensor") + sys.exit(1) + + if base_name in tensor_map: + if is_lora_a: + tensor_map[base_name].A = tensor + else: + tensor_map[base_name].B = tensor + else: + if is_lora_a: + tensor_map[base_name] = PartialLoraTensor(A=tensor) + else: + tensor_map[base_name] = PartialLoraTensor(B=tensor) + + for name, tensor in tensor_map.items(): + assert tensor.A is not None + assert tensor.B is not None + yield (name, cast(torch.Tensor, LoraTorchTensor(tensor.A, tensor.B))) + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + dest = super().modify_tensors(data_torch, name, bid) + for dest_name, dest_data in dest: + assert isinstance(dest_data, LoraTorchTensor) + lora_a, lora_b = dest_data.get_lora_A_B() + + yield (dest_name + ".lora_a", lora_a) + yield (dest_name + ".lora_b", lora_b) + + with open(lora_config, "r") as f: + lparams: dict[str, Any] = json.load(f) + + alpha: float = lparams["lora_alpha"] + + model_instance = LoraModel( + dir_base_model, + ftype, + fname_out, + is_big_endian=args.bigendian, + use_temp_file=False, + eager=args.no_lazy, + dry_run=args.dry_run, + dir_lora_model=dir_lora, + lora_alpha=alpha, + ) + + logger.info("Exporting model...") + model_instance.write() + logger.info(f"Model successfully exported to {model_instance.fname_out}") diff --git a/docs/android.md b/docs/android.md new file mode 100644 index 000000000..cec4358d9 --- /dev/null +++ b/docs/android.md @@ -0,0 +1,56 @@ + +# Android + +## Build on Android using Termux +[Termux](https://github.com/termux/termux-app#installation) is a method to execute `llama.cpp` on an Android device (no root required). +``` +apt update && apt upgrade -y +apt install git make cmake +``` + +It's recommended to move your model inside the `~/` directory for best performance: +``` +cd storage/downloads +mv model.gguf ~/ +``` + +[Get the code](https://github.com/ggerganov/llama.cpp#get-the-code) & [follow the Linux build instructions](https://github.com/ggerganov/llama.cpp#build) to build `llama.cpp`. + +## Building the Project using Android NDK +Obtain the [Android NDK](https://developer.android.com/ndk) and then build with CMake. + +Execute the following commands on your computer to avoid downloading the NDK to your mobile. Alternatively, you can also do this in Termux: +``` +$ mkdir build-android +$ cd build-android +$ export NDK= +$ cmake -DCMAKE_TOOLCHAIN_FILE=$NDK/build/cmake/android.toolchain.cmake -DANDROID_ABI=arm64-v8a -DANDROID_PLATFORM=android-23 -DCMAKE_C_FLAGS=-march=armv8.4a+dotprod .. +$ make +``` + +Install [termux](https://github.com/termux/termux-app#installation) on your device and run `termux-setup-storage` to get access to your SD card (if Android 11+ then run the command twice). + +Finally, copy these built `llama` binaries and the model file to your device storage. Because the file permissions in the Android sdcard cannot be changed, you can copy the executable files to the `/data/data/com.termux/files/home/bin` path, and then execute the following commands in Termux to add executable permission: + +(Assumed that you have pushed the built executable files to the /sdcard/llama.cpp/bin path using `adb push`) +``` +$cp -r /sdcard/llama.cpp/bin /data/data/com.termux/files/home/ +$cd /data/data/com.termux/files/home/bin +$chmod +x ./* +``` + +Download model [llama-2-7b-chat.Q4_K_M.gguf](https://huggingface.co/TheBloke/Llama-2-7B-Chat-GGUF/blob/main/llama-2-7b-chat.Q4_K_M.gguf), and push it to `/sdcard/llama.cpp/`, then move it to `/data/data/com.termux/files/home/model/` + +``` +$mv /sdcard/llama.cpp/llama-2-7b-chat.Q4_K_M.gguf /data/data/com.termux/files/home/model/ +``` + +Now, you can start chatting: +``` +$cd /data/data/com.termux/files/home/bin +$./llama-cli -m ../model/llama-2-7b-chat.Q4_K_M.gguf -n 128 -cml +``` + +Here's a demo of an interactive session running on Pixel 5 phone: + +https://user-images.githubusercontent.com/271616/225014776-1d567049-ad71-4ef2-b050-55b0b3b9274c.mp4 diff --git a/docs/BLIS.md b/docs/backend/BLIS.md similarity index 100% rename from docs/BLIS.md rename to docs/backend/BLIS.md diff --git a/README-sycl.md b/docs/backend/SYCL.md similarity index 91% rename from README-sycl.md rename to docs/backend/SYCL.md index 885983e92..d36ac0a15 100644 --- a/README-sycl.md +++ b/docs/backend/SYCL.md @@ -293,31 +293,26 @@ Similar to the native `sycl-ls`, available SYCL devices can be queried as follow ```sh ./build/bin/llama-ls-sycl-device ``` -A example of such log in a system with 1 *intel CPU* and 1 *intel GPU* can look like the following: +This command will only display the selected backend that is supported by SYCL. The default backend is level_zero. For example, in a system with 2 *intel GPU* it would look like the following: ``` -found 6 SYCL devices: +found 2 SYCL devices: + | | | |Compute |Max compute|Max work|Max sub| | |ID| Device Type| Name|capability|units |group |group |Global mem size| |--|------------------|---------------------------------------------|----------|-----------|--------|-------|---------------| | 0|[level_zero:gpu:0]| Intel(R) Arc(TM) A770 Graphics| 1.3| 512| 1024| 32| 16225243136| | 1|[level_zero:gpu:1]| Intel(R) UHD Graphics 770| 1.3| 32| 512| 32| 53651849216| -| 2| [opencl:gpu:0]| Intel(R) Arc(TM) A770 Graphics| 3.0| 512| 1024| 32| 16225243136| -| 3| [opencl:gpu:1]| Intel(R) UHD Graphics 770| 3.0| 32| 512| 32| 53651849216| -| 4| [opencl:cpu:0]| 13th Gen Intel(R) Core(TM) i7-13700K| 3.0| 24| 8192| 64| 67064815616| -| 5| [opencl:acc:0]| Intel(R) FPGA Emulation Device| 1.2| 24|67108864| 64| 67064815616| ``` -| Attribute | Note | -|------------------------|-------------------------------------------------------------| -| compute capability 1.3 | Level-zero driver/runtime, recommended | -| compute capability 3.0 | OpenCL driver/runtime, slower than level-zero in most cases | 4. Launch inference There are two device selection modes: - Single device: Use one device target specified by the user. -- Multiple devices: Automatically select the devices with the same largest Max compute-units. +- Multiple devices: Automatically choose the devices with the same backend. + +In two device selection modes, the default SYCL backend is level_zero, you can choose other backend supported by SYCL by setting environment variable ONEAPI_DEVICE_SELECTOR. | Device selection | Parameter | |------------------|----------------------------------------| @@ -474,33 +469,26 @@ Similar to the native `sycl-ls`, available SYCL devices can be queried as follow build\bin\ls-sycl-device.exe ``` -The output of this command in a system with 1 *intel CPU* and 1 *intel GPU* would look like the following: +This command will only display the selected backend that is supported by SYCL. The default backend is level_zero. For example, in a system with 2 *intel GPU* it would look like the following: ``` -found 6 SYCL devices: +found 2 SYCL devices: | | | |Compute |Max compute|Max work|Max sub| | |ID| Device Type| Name|capability|units |group |group |Global mem size| |--|------------------|---------------------------------------------|----------|-----------|--------|-------|---------------| | 0|[level_zero:gpu:0]| Intel(R) Arc(TM) A770 Graphics| 1.3| 512| 1024| 32| 16225243136| | 1|[level_zero:gpu:1]| Intel(R) UHD Graphics 770| 1.3| 32| 512| 32| 53651849216| -| 2| [opencl:gpu:0]| Intel(R) Arc(TM) A770 Graphics| 3.0| 512| 1024| 32| 16225243136| -| 3| [opencl:gpu:1]| Intel(R) UHD Graphics 770| 3.0| 32| 512| 32| 53651849216| -| 4| [opencl:cpu:0]| 13th Gen Intel(R) Core(TM) i7-13700K| 3.0| 24| 8192| 64| 67064815616| -| 5| [opencl:acc:0]| Intel(R) FPGA Emulation Device| 1.2| 24|67108864| 64| 67064815616| ``` -| Attribute | Note | -|------------------------|-----------------------------------------------------------| -| compute capability 1.3 | Level-zero running time, recommended | -| compute capability 3.0 | OpenCL running time, slower than level-zero in most cases | - 4. Launch inference There are two device selection modes: -- Single device: Use one device assigned by user. -- Multiple devices: Automatically choose the devices with the same biggest Max compute units. +- Single device: Use one device assigned by user. Default device id is 0. +- Multiple devices: Automatically choose the devices with the same backend. + +In two device selection modes, the default SYCL backend is level_zero, you can choose other backend supported by SYCL by setting environment variable ONEAPI_DEVICE_SELECTOR. | Device selection | Parameter | |------------------|----------------------------------------| diff --git a/docs/build.md b/docs/build.md new file mode 100644 index 000000000..cfe42ebbf --- /dev/null +++ b/docs/build.md @@ -0,0 +1,353 @@ +# Build llama.cpp locally + +**To get the Code:** + +```bash +git clone https://github.com/ggerganov/llama.cpp +cd llama.cpp +``` + +In order to build llama.cpp you have four different options. + +- Using `make`: + - On Linux or MacOS: + + ```bash + make + ``` + + - On Windows (x86/x64 only, arm64 requires cmake): + + 1. Download the latest fortran version of [w64devkit](https://github.com/skeeto/w64devkit/releases). + 2. Extract `w64devkit` on your pc. + 3. Run `w64devkit.exe`. + 4. Use the `cd` command to reach the `llama.cpp` folder. + 5. From here you can run: + ```bash + make + ``` + + - Notes: + - For `Q4_0_4_4` quantization type build, add the `GGML_NO_LLAMAFILE=1` flag. For example, use `make GGML_NO_LLAMAFILE=1`. + - For faster compilation, add the `-j` argument to run multiple jobs in parallel. For example, `make -j 8` will run 8 jobs in parallel. + - For faster repeated compilation, install [ccache](https://ccache.dev/). + - For debug builds, run `make LLAMA_DEBUG=1` + +- Using `CMake`: + + ```bash + cmake -B build + cmake --build build --config Release + ``` + + **Notes**: + + - For `Q4_0_4_4` quantization type build, add the `-DGGML_LLAMAFILE=OFF` cmake option. For example, use `cmake -B build -DGGML_LLAMAFILE=OFF`. + - For faster compilation, add the `-j` argument to run multiple jobs in parallel. For example, `cmake --build build --config Release -j 8` will run 8 jobs in parallel. + - For faster repeated compilation, install [ccache](https://ccache.dev/). + - For debug builds, there are two cases: + + 1. Single-config generators (e.g. default = `Unix Makefiles`; note that they just ignore the `--config` flag): + + ```bash + cmake -B build -DCMAKE_BUILD_TYPE=Debug + cmake --build build + ``` + + 2. Multi-config generators (`-G` param set to Visual Studio, XCode...): + + ```bash + cmake -B build -G "Xcode" + cmake --build build --config Debug + ``` + - Building for Windows (x86, x64 and arm64) with MSVC or clang as compilers: + - Install Visual Studio 2022, e.g. via the [Community Edition](https://visualstudio.microsoft.com/de/vs/community/). In the installer, select at least the following options (this also automatically installs the required additional tools like CMake,...): + - Tab Workload: Desktop-development with C++ + - Tab Components (select quickly via search): C++-_CMake_ Tools for Windows, _Git_ for Windows, C++-_Clang_ Compiler for Windows, MS-Build Support for LLVM-Toolset (clang) + - Please remember to always use a Developer Command Prompt / PowerShell for VS2022 for git, build, test + - For Windows on ARM (arm64, WoA) build with: + ```bash + cmake --preset arm64-windows-llvm-release -D GGML_OPENMP=OFF + cmake --build build-arm64-windows-llvm-release + ``` + Note: Building for arm64 could also be done just with MSVC (with the build-arm64-windows-MSVC preset, or the standard CMake build instructions). But MSVC does not support inline ARM assembly-code, used e.g. for the accelerated Q4_0_4_8 CPU kernels. + +- Using `gmake` (FreeBSD): + + 1. Install and activate [DRM in FreeBSD](https://wiki.freebsd.org/Graphics) + 2. Add your user to **video** group + 3. Install compilation dependencies. + + ```bash + sudo pkg install gmake automake autoconf pkgconf llvm15 openblas + + gmake CC=/usr/local/bin/clang15 CXX=/usr/local/bin/clang++15 -j4 + ``` + +## Metal Build + +On MacOS, Metal is enabled by default. Using Metal makes the computation run on the GPU. +To disable the Metal build at compile time use the `GGML_NO_METAL=1` flag or the `GGML_METAL=OFF` cmake option. + +When built with Metal support, you can explicitly disable GPU inference with the `--n-gpu-layers|-ngl 0` command-line +argument. + +## BLAS Build + +Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). Support with CPU-only BLAS implementations doesn't affect the normal generation performance. We may see generation performance improvements with GPU-involved BLAS implementations, e.g. cuBLAS, hipBLAS. There are currently several different BLAS implementations available for build and use: + +### Accelerate Framework: + +This is only available on Mac PCs and it's enabled by default. You can just build using the normal instructions. + +### OpenBLAS: + +This provides BLAS acceleration using only the CPU. Make sure to have OpenBLAS installed on your machine. + +- Using `make`: + - On Linux: + ```bash + make GGML_OPENBLAS=1 + ``` + + - On Windows: + + 1. Download the latest fortran version of [w64devkit](https://github.com/skeeto/w64devkit/releases). + 2. Download the latest version of [OpenBLAS for Windows](https://github.com/xianyi/OpenBLAS/releases). + 3. Extract `w64devkit` on your pc. + 4. From the OpenBLAS zip that you just downloaded copy `libopenblas.a`, located inside the `lib` folder, inside `w64devkit\x86_64-w64-mingw32\lib`. + 5. From the same OpenBLAS zip copy the content of the `include` folder inside `w64devkit\x86_64-w64-mingw32\include`. + 6. Run `w64devkit.exe`. + 7. Use the `cd` command to reach the `llama.cpp` folder. + 8. From here you can run: + + ```bash + make GGML_OPENBLAS=1 + ``` + +- Using `CMake` on Linux: + + ```bash + cmake -B build -DGGML_BLAS=ON -DGGML_BLAS_VENDOR=OpenBLAS + cmake --build build --config Release + ``` + +### BLIS + +Check [BLIS.md](./backend/BLIS.md) for more information. + +### SYCL + +SYCL is a higher-level programming model to improve programming productivity on various hardware accelerators. + +llama.cpp based on SYCL is used to **support Intel GPU** (Data Center Max series, Flex series, Arc series, Built-in GPU and iGPU). + +For detailed info, please refer to [llama.cpp for SYCL](./backend/SYCL.md). + +### Intel oneMKL + +Building through oneAPI compilers will make avx_vnni instruction set available for intel processors that do not support avx512 and avx512_vnni. Please note that this build config **does not support Intel GPU**. For Intel GPU support, please refer to [llama.cpp for SYCL](./backend/SYCL.md). + +- Using manual oneAPI installation: + By default, `GGML_BLAS_VENDOR` is set to `Generic`, so if you already sourced intel environment script and assign `-DGGML_BLAS=ON` in cmake, the mkl version of Blas will automatically been selected. Otherwise please install oneAPI and follow the below steps: + ```bash + source /opt/intel/oneapi/setvars.sh # You can skip this step if in oneapi-basekit docker image, only required for manual installation + cmake -B build -DGGML_BLAS=ON -DGGML_BLAS_VENDOR=Intel10_64lp -DCMAKE_C_COMPILER=icx -DCMAKE_CXX_COMPILER=icpx -DGGML_NATIVE=ON + cmake --build build --config Release + ``` + +- Using oneAPI docker image: + If you do not want to source the environment vars and install oneAPI manually, you can also build the code using intel docker container: [oneAPI-basekit](https://hub.docker.com/r/intel/oneapi-basekit). Then, you can use the commands given above. + +Check [Optimizing and Running LLaMA2 on Intel® CPU](https://www.intel.com/content/www/us/en/content-details/791610/optimizing-and-running-llama2-on-intel-cpu.html) for more information. + +### CUDA + +This provides GPU acceleration using the CUDA cores of your Nvidia GPU. Make sure to have the CUDA toolkit installed. You can download it from your Linux distro's package manager (e.g. `apt install nvidia-cuda-toolkit`) or from here: [CUDA Toolkit](https://developer.nvidia.com/cuda-downloads). + +For Jetson user, if you have Jetson Orin, you can try this: [Offical Support](https://www.jetson-ai-lab.com/tutorial_text-generation.html). If you are using an old model(nano/TX2), need some additional operations before compiling. + +- Using `make`: + ```bash + make GGML_CUDA=1 + ``` +- Using `CMake`: + + ```bash + cmake -B build -DGGML_CUDA=ON + cmake --build build --config Release + ``` + +The environment variable [`CUDA_VISIBLE_DEVICES`](https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars) can be used to specify which GPU(s) will be used. The following compilation options are also available to tweak performance: + +| Option | Legal values | Default | Description | +|-------------------------------|------------------------|---------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| GGML_CUDA_FORCE_DMMV | Boolean | false | Force the use of dequantization + matrix vector multiplication kernels instead of using kernels that do matrix vector multiplication on quantized data. By default the decision is made based on compute capability (MMVQ for 6.1/Pascal/GTX 1000 or higher). Does not affect k-quants. | +| GGML_CUDA_DMMV_X | Positive integer >= 32 | 32 | Number of values in x direction processed by the CUDA dequantization + matrix vector multiplication kernel per iteration. Increasing this value can improve performance on fast GPUs. Power of 2 heavily recommended. Does not affect k-quants. | +| GGML_CUDA_MMV_Y | Positive integer | 1 | Block size in y direction for the CUDA mul mat vec kernels. Increasing this value can improve performance on fast GPUs. Power of 2 recommended. | +| GGML_CUDA_FORCE_MMQ | Boolean | false | Force the use of custom matrix multiplication kernels for quantized models instead of FP16 cuBLAS even if there is no int8 tensor core implementation available (affects V100, RDNA3). MMQ kernels are enabled by default on GPUs with int8 tensor core support. With MMQ force enabled, speed for large batch sizes will be worse but VRAM consumption will be lower. | +| GGML_CUDA_FORCE_CUBLAS | Boolean | false | Force the use of FP16 cuBLAS instead of custom matrix multiplication kernels for quantized models | +| GGML_CUDA_F16 | Boolean | false | If enabled, use half-precision floating point arithmetic for the CUDA dequantization + mul mat vec kernels and for the q4_1 and q5_1 matrix matrix multiplication kernels. Can improve performance on relatively recent GPUs. | +| GGML_CUDA_KQUANTS_ITER | 1 or 2 | 2 | Number of values processed per iteration and per CUDA thread for Q2_K and Q6_K quantization formats. Setting this value to 1 can improve performance for slow GPUs. | +| GGML_CUDA_PEER_MAX_BATCH_SIZE | Positive integer | 128 | Maximum batch size for which to enable peer access between multiple GPUs. Peer access requires either Linux or NVLink. When using NVLink enabling peer access for larger batch sizes is potentially beneficial. | +| GGML_CUDA_FA_ALL_QUANTS | Boolean | false | Compile support for all KV cache quantization type (combinations) for the FlashAttention CUDA kernels. More fine-grained control over KV cache size but compilation takes much longer. | + +### MUSA + +- Using `make`: + ```bash + make GGML_MUSA=1 + ``` +- Using `CMake`: + + ```bash + cmake -B build -DGGML_MUSA=ON + cmake --build build --config Release + ``` + +### hipBLAS + +This provides BLAS acceleration on HIP-supported AMD GPUs. +Make sure to have ROCm installed. +You can download it from your Linux distro's package manager or from here: [ROCm Quick Start (Linux)](https://rocm.docs.amd.com/projects/install-on-linux/en/latest/tutorial/quick-start.html#rocm-install-quick). + +- Using `make`: + ```bash + make GGML_HIPBLAS=1 + ``` +- Using `CMake` for Linux (assuming a gfx1030-compatible AMD GPU): + ```bash + HIPCXX="$(hipconfig -l)/clang" HIP_PATH="$(hipconfig -R)" \ + cmake -S . -B build -DGGML_HIPBLAS=ON -DAMDGPU_TARGETS=gfx1030 -DCMAKE_BUILD_TYPE=Release \ + && cmake --build build --config Release -- -j 16 + ``` + On Linux it is also possible to use unified memory architecture (UMA) to share main memory between the CPU and integrated GPU by setting `-DGGML_HIP_UMA=ON`. + However, this hurts performance for non-integrated GPUs (but enables working with integrated GPUs). + + Note that if you get the following error: + ``` + clang: error: cannot find ROCm device library; provide its path via '--rocm-path' or '--rocm-device-lib-path', or pass '-nogpulib' to build without ROCm device library + ``` + Try searching for a directory under `HIP_PATH` that contains the file + `oclc_abi_version_400.bc`. Then, add the following to the start of the + command: `HIP_DEVICE_LIB_PATH=`, so something + like: + ```bash + HIPCXX="$(hipconfig -l)/clang" HIP_PATH="$(hipconfig -p)" \ + HIP_DEVICE_LIB_PATH= \ + cmake -S . -B build -DGGML_HIPBLAS=ON -DAMDGPU_TARGETS=gfx1030 -DCMAKE_BUILD_TYPE=Release \ + && cmake --build build -- -j 16 + ``` + +- Using `make` (example for target gfx1030, build with 16 CPU threads): + ```bash + make -j16 GGML_HIPBLAS=1 GGML_HIP_UMA=1 AMDGPU_TARGETS=gfx1030 + ``` + +- Using `CMake` for Windows (using x64 Native Tools Command Prompt for VS, and assuming a gfx1100-compatible AMD GPU): + ```bash + set PATH=%HIP_PATH%\bin;%PATH% + cmake -S . -B build -G Ninja -DAMDGPU_TARGETS=gfx1100 -DGGML_HIPBLAS=ON -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_BUILD_TYPE=Release + cmake --build build + ``` + Make sure that `AMDGPU_TARGETS` is set to the GPU arch you want to compile for. The above example uses `gfx1100` that corresponds to Radeon RX 7900XTX/XT/GRE. You can find a list of targets [here](https://llvm.org/docs/AMDGPUUsage.html#processors) + Find your gpu version string by matching the most significant version information from `rocminfo | grep gfx | head -1 | awk '{print $2}'` with the list of processors, e.g. `gfx1035` maps to `gfx1030`. + + +The environment variable [`HIP_VISIBLE_DEVICES`](https://rocm.docs.amd.com/en/latest/understand/gpu_isolation.html#hip-visible-devices) can be used to specify which GPU(s) will be used. +If your GPU is not officially supported you can use the environment variable [`HSA_OVERRIDE_GFX_VERSION`] set to a similar GPU, for example 10.3.0 on RDNA2 (e.g. gfx1030, gfx1031, or gfx1035) or 11.0.0 on RDNA3. +The following compilation options are also available to tweak performance (yes, they refer to CUDA, not HIP, because it uses the same code as the cuBLAS version above): + +| Option | Legal values | Default | Description | +|------------------------|------------------------|---------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| +| GGML_CUDA_DMMV_X | Positive integer >= 32 | 32 | Number of values in x direction processed by the HIP dequantization + matrix vector multiplication kernel per iteration. Increasing this value can improve performance on fast GPUs. Power of 2 heavily recommended. Does not affect k-quants. | +| GGML_CUDA_MMV_Y | Positive integer | 1 | Block size in y direction for the HIP mul mat vec kernels. Increasing this value can improve performance on fast GPUs. Power of 2 recommended. Does not affect k-quants. | +| GGML_CUDA_KQUANTS_ITER | 1 or 2 | 2 | Number of values processed per iteration and per HIP thread for Q2_K and Q6_K quantization formats. Setting this value to 1 can improve performance for slow GPUs. | + +### Vulkan + +**Windows** + +#### w64devkit + +Download and extract [w64devkit](https://github.com/skeeto/w64devkit/releases). + +Download and install the [Vulkan SDK](https://vulkan.lunarg.com/sdk/home#windows). When selecting components, only the Vulkan SDK Core is required. + +Launch `w64devkit.exe` and run the following commands to copy Vulkan dependencies: +```sh +SDK_VERSION=1.3.283.0 +cp /VulkanSDK/$SDK_VERSION/Bin/glslc.exe $W64DEVKIT_HOME/bin/ +cp /VulkanSDK/$SDK_VERSION/Lib/vulkan-1.lib $W64DEVKIT_HOME/x86_64-w64-mingw32/lib/ +cp -r /VulkanSDK/$SDK_VERSION/Include/* $W64DEVKIT_HOME/x86_64-w64-mingw32/include/ +cat > $W64DEVKIT_HOME/x86_64-w64-mingw32/lib/pkgconfig/vulkan.pc <' --install --attr llama-cpp +``` + +For non-flake enabled installs. + +This expression is automatically updated within the [nixpkgs repo](https://github.com/NixOS/nixpkgs/blob/nixos-24.05/pkgs/by-name/ll/llama-cpp/package.nix#L164). + +## Flox + +On Mac and Linux, Flox can be used to install llama.cpp within a Flox environment via + +```sh +flox install llama-cpp +``` + +Flox follows the nixpkgs build of llama.cpp. diff --git a/examples/CMakeLists.txt b/examples/CMakeLists.txt index 7d9ab3457..67b3d2774 100644 --- a/examples/CMakeLists.txt +++ b/examples/CMakeLists.txt @@ -21,8 +21,8 @@ else() add_subdirectory(embedding) add_subdirectory(eval-callback) add_subdirectory(export-lora) - add_subdirectory(finetune) add_subdirectory(gbnf-validator) + add_subdirectory(gguf-hash) add_subdirectory(gguf-split) add_subdirectory(gguf) add_subdirectory(gritlm) @@ -52,5 +52,4 @@ else() add_subdirectory(simple) add_subdirectory(speculative) add_subdirectory(tokenize) - add_subdirectory(train-text-from-scratch) endif() diff --git a/examples/batched.swift/Sources/main.swift b/examples/batched.swift/Sources/main.swift index dbbd06da5..616494d2d 100644 --- a/examples/batched.swift/Sources/main.swift +++ b/examples/batched.swift/Sources/main.swift @@ -229,7 +229,7 @@ private func tokenize(text: String, add_bos: Bool) -> [llama_token] { private func token_to_piece(token: llama_token, buffer: inout [CChar]) -> String? { var result = [CChar](repeating: 0, count: 8) - let nTokens = llama_token_to_piece(model, token, &result, Int32(result.count), false) + let nTokens = llama_token_to_piece(model, token, &result, Int32(result.count), 0, false) if nTokens < 0 { let actualTokensCount = -Int(nTokens) result = .init(repeating: 0, count: actualTokensCount) @@ -238,6 +238,7 @@ private func token_to_piece(token: llama_token, buffer: inout [CChar]) -> String token, &result, Int32(result.count), + 0, false ) assert(check == actualTokensCount) diff --git a/examples/batched/batched.cpp b/examples/batched/batched.cpp index 62d9b144d..53fbfb0a8 100644 --- a/examples/batched/batched.cpp +++ b/examples/batched/batched.cpp @@ -31,7 +31,7 @@ int main(int argc, char ** argv) { int n_parallel = params.n_parallel; // total length of the sequences including the prompt - int n_predict = 32; + int n_predict = params.n_predict; // init LLM @@ -93,14 +93,34 @@ int main(int argc, char ** argv) { // create a llama_batch // we use this object to submit token data for decoding - llama_batch batch = llama_batch_init(std::max(tokens_list.size(), (size_t)n_parallel), 0, 1); + llama_batch batch = llama_batch_init(std::max(tokens_list.size(), (size_t) n_parallel), 0, n_parallel); + + std::vector seq_ids(n_parallel, 0); + for (int32_t i = 0; i < n_parallel; ++i) { + seq_ids[i] = i; + } // evaluate the initial prompt for (size_t i = 0; i < tokens_list.size(); ++i) { - llama_batch_add(batch, tokens_list[i], i, { 0 }, false); + llama_batch_add(batch, tokens_list[i], i, seq_ids, false); } GGML_ASSERT(batch.n_tokens == (int) tokens_list.size()); + if (llama_model_has_encoder(model)) { + if (llama_encode(ctx, batch)) { + LOG_TEE("%s : failed to eval\n", __func__); + return 1; + } + + llama_token decoder_start_token_id = llama_model_decoder_start_token(model); + if (decoder_start_token_id == -1) { + decoder_start_token_id = llama_token_bos(model); + } + + llama_batch_clear(batch); + llama_batch_add(batch, decoder_start_token_id, 0, seq_ids, false); + } + // llama_decode will output logits only for the last token of the prompt batch.logits[batch.n_tokens - 1] = true; @@ -109,11 +129,11 @@ int main(int argc, char ** argv) { return 1; } - // assign the system KV cache to all parallel sequences - // this way, the parallel sequences will "reuse" the prompt tokens without having to copy them - for (int32_t i = 1; i < n_parallel; ++i) { - llama_kv_cache_seq_cp(ctx, 0, i, -1, -1); - } + //// assign the system KV cache to all parallel sequences + //// this way, the parallel sequences will "reuse" the prompt tokens without having to copy them + //for (int32_t i = 1; i < n_parallel; ++i) { + // llama_kv_cache_seq_cp(ctx, 0, i, -1, -1); + //} if (n_parallel > 1) { LOG_TEE("\n\n%s: generating %d sequences ...\n", __func__, n_parallel); diff --git a/examples/convert-legacy-llama.py b/examples/convert_legacy_llama.py similarity index 88% rename from examples/convert-legacy-llama.py rename to examples/convert_legacy_llama.py index 721a57c00..9ab9ab06e 100755 --- a/examples/convert-legacy-llama.py +++ b/examples/convert_legacy_llama.py @@ -24,7 +24,7 @@ from abc import ABC, abstractmethod from concurrent.futures import ProcessPoolExecutor, ThreadPoolExecutor from dataclasses import dataclass from pathlib import Path -from typing import TYPE_CHECKING, Any, Callable, IO, Iterable, Literal, TypeVar, Optional +from typing import TYPE_CHECKING, Any, Callable, IO, Iterable, Literal, TypeVar import numpy as np @@ -346,42 +346,6 @@ class Params: return params -@dataclass -class Metadata: - name: Optional[str] = None - author: Optional[str] = None - version: Optional[str] = None - url: Optional[str] = None - description: Optional[str] = None - licence: Optional[str] = None - source_url: Optional[str] = None - source_hf_repo: Optional[str] = None - - @staticmethod - def load(metadata_path: Path) -> Metadata: - if metadata_path is None or not metadata_path.exists(): - return Metadata() - - with open(metadata_path, 'r') as file: - data = json.load(file) - - # Create a new Metadata instance - metadata = Metadata() - - # Assigning values to Metadata attributes if they exist in the JSON file - # This is based on LLM_KV_NAMES mapping in llama.cpp - metadata.name = data.get("general.name") - metadata.author = data.get("general.author") - metadata.version = data.get("general.version") - metadata.url = data.get("general.url") - metadata.description = data.get("general.description") - metadata.license = data.get("general.license") - metadata.source_url = data.get("general.source.url") - metadata.source_hf_repo = data.get("general.source.huggingface.repository") - - return metadata - - # # data loading # TODO: reuse (probably move to gguf.py?) @@ -492,12 +456,13 @@ class LazyTensor: LazyModel: TypeAlias = 'dict[str, LazyTensor]' +ModelFormat: TypeAlias = Literal['ggml', 'torch', 'safetensors', 'none'] @dataclass class ModelPlus: model: LazyModel paths: list[Path] # Where this was read from. - format: Literal['ggml', 'torch', 'safetensors', 'none'] + format: ModelFormat vocab: BaseVocab | None # For GGML models (which have vocab built in), the vocab. @@ -536,7 +501,7 @@ def merge_sharded(models: list[LazyModel]) -> LazyModel: def merge_multifile_models(models_plus: list[ModelPlus]) -> ModelPlus: - formats = set(mp.format for mp in models_plus) + formats: set[ModelFormat] = set(mp.format for mp in models_plus) assert len(formats) == 1, "different formats?" format = formats.pop() paths = [path for mp in models_plus for path in mp.paths] @@ -555,7 +520,7 @@ def merge_multifile_models(models_plus: list[ModelPlus]) -> ModelPlus: else: model = merge_sharded([mp.model for mp in models_plus]) - return ModelPlus(model, paths, format, vocab) # pytype: disable=wrong-arg-types + return ModelPlus(model, paths, format, vocab) def permute_lazy(lazy_tensor: LazyTensor, n_head: int, n_head_kv: int) -> LazyTensor: @@ -805,7 +770,7 @@ class OutputFile: def __init__(self, fname_out: Path, endianess:gguf.GGUFEndian = gguf.GGUFEndian.LITTLE): self.gguf = gguf.GGUFWriter(fname_out, gguf.MODEL_ARCH_NAMES[ARCH], endianess=endianess) - def add_meta_model(self, params: Params, metadata: Metadata) -> None: + def add_meta_model(self, params: Params, metadata: gguf.Metadata | None) -> None: # Metadata About The Model And Its Provenence name = "LLaMA" if metadata is not None and metadata.name is not None: @@ -823,16 +788,73 @@ class OutputFile: self.gguf.add_author(metadata.author) if metadata.version is not None: self.gguf.add_version(metadata.version) - if metadata.url is not None: - self.gguf.add_url(metadata.url) + if metadata.organization is not None: + self.gguf.add_organization(metadata.organization) + + if metadata.finetune is not None: + self.gguf.add_finetune(metadata.finetune) + if metadata.basename is not None: + self.gguf.add_basename(metadata.basename) + if metadata.description is not None: self.gguf.add_description(metadata.description) - if metadata.licence is not None: - self.gguf.add_licence(metadata.licence) + if metadata.quantized_by is not None: + self.gguf.add_quantized_by(metadata.quantized_by) + + if metadata.size_label is not None: + self.gguf.add_size_label(metadata.size_label) + + if metadata.license is not None: + self.gguf.add_license(metadata.license) + if metadata.license_name is not None: + self.gguf.add_license_name(metadata.license_name) + if metadata.license_link is not None: + self.gguf.add_license_link(metadata.license_link) + + if metadata.url is not None: + self.gguf.add_url(metadata.url) + if metadata.doi is not None: + self.gguf.add_doi(metadata.doi) + if metadata.uuid is not None: + self.gguf.add_uuid(metadata.uuid) + if metadata.repo_url is not None: + self.gguf.add_repo_url(metadata.repo_url) + if metadata.source_url is not None: self.gguf.add_source_url(metadata.source_url) - if metadata.source_hf_repo is not None: - self.gguf.add_source_hf_repo(metadata.source_hf_repo) + if metadata.source_doi is not None: + self.gguf.add_source_doi(metadata.source_doi) + if metadata.source_uuid is not None: + self.gguf.add_source_uuid(metadata.source_uuid) + if metadata.source_repo_url is not None: + self.gguf.add_source_repo_url(metadata.source_repo_url) + + if metadata.base_models is not None: + self.gguf.add_base_model_count(len(metadata.base_models)) + for key, base_model_entry in enumerate(metadata.base_models): + if "name" in base_model_entry: + self.gguf.add_base_model_name(key, base_model_entry["name"]) + if "author" in base_model_entry: + self.gguf.add_base_model_author(key, base_model_entry["author"]) + if "version" in base_model_entry: + self.gguf.add_base_model_version(key, base_model_entry["version"]) + if "organization" in base_model_entry: + self.gguf.add_base_model_organization(key, base_model_entry["organization"]) + if "url" in base_model_entry: + self.gguf.add_base_model_url(key, base_model_entry["url"]) + if "doi" in base_model_entry: + self.gguf.add_base_model_doi(key, base_model_entry["doi"]) + if "uuid" in base_model_entry: + self.gguf.add_base_model_uuid(key, base_model_entry["uuid"]) + if "repo_url" in base_model_entry: + self.gguf.add_base_model_repo_url(key, base_model_entry["repo_url"]) + + if metadata.tags is not None: + self.gguf.add_tags(metadata.tags) + if metadata.languages is not None: + self.gguf.add_languages(metadata.languages) + if metadata.datasets is not None: + self.gguf.add_datasets(metadata.datasets) def add_meta_arch(self, params: Params) -> None: # Metadata About The Neural Architecture Itself @@ -943,7 +965,7 @@ class OutputFile: @staticmethod def write_vocab_only( fname_out: Path, params: Params, vocab: Vocab, svocab: gguf.SpecialVocab, - endianess: gguf.GGUFEndian = gguf.GGUFEndian.LITTLE, pad_vocab: bool = False, metadata: Metadata = None, + endianess: gguf.GGUFEndian = gguf.GGUFEndian.LITTLE, pad_vocab: bool = False, metadata: gguf.Metadata | None = None, ) -> None: check_vocab_size(params, vocab, pad_vocab=pad_vocab) @@ -977,7 +999,7 @@ class OutputFile: fname_out: Path, ftype: GGMLFileType, params: Params, model: LazyModel, vocab: BaseVocab, svocab: gguf.SpecialVocab, concurrency: int = DEFAULT_CONCURRENCY, endianess: gguf.GGUFEndian = gguf.GGUFEndian.LITTLE, pad_vocab: bool = False, - metadata: Metadata = None, + metadata: gguf.Metadata | None = None, ) -> None: check_vocab_size(params, vocab, pad_vocab=pad_vocab) @@ -1020,35 +1042,32 @@ def pick_output_type(model: LazyModel, output_type_str: str | None) -> GGMLFileT raise ValueError(f"Unexpected combination of types: {name_to_type}") -def model_parameter_count(model: LazyModel) -> int: - total_model_parameters = 0 - for i, (name, lazy_tensor) in enumerate(model.items()): - sum_weights_in_tensor = 1 +def per_model_weight_count_estimation(tensors: Iterable[tuple[str, LazyTensor]]) -> tuple[int, int, int]: + total_params = 0 + shared_params = 0 + expert_params = 0 + + for name, lazy_tensor in tensors: + # We don't need these + if name.endswith((".attention.masked_bias", ".attention.bias", ".rotary_emb.inv_freq")): + continue + + # Got A Tensor + sum_weights_in_tensor: int = 1 + + # Tensor Volume for dim in lazy_tensor.shape: sum_weights_in_tensor *= dim - total_model_parameters += sum_weights_in_tensor - return total_model_parameters + if ".experts." in name: + if ".experts.0." in name: + expert_params += sum_weights_in_tensor + else: + shared_params += sum_weights_in_tensor -def model_parameter_count_rounded_notation(model_params_count: int) -> str: - if model_params_count > 1e12 : - # Trillions Of Parameters - scaled_model_params = model_params_count * 1e-12 - scale_suffix = "T" - elif model_params_count > 1e9 : - # Billions Of Parameters - scaled_model_params = model_params_count * 1e-9 - scale_suffix = "B" - elif model_params_count > 1e6 : - # Millions Of Parameters - scaled_model_params = model_params_count * 1e-6 - scale_suffix = "M" - else: - # Thousands Of Parameters - scaled_model_params = model_params_count * 1e-3 - scale_suffix = "K" + total_params += sum_weights_in_tensor - return f"{round(scaled_model_params)}{scale_suffix}" + return total_params, shared_params, expert_params def convert_to_output_type(model: LazyModel, output_type: GGMLFileType) -> LazyModel: @@ -1230,34 +1249,24 @@ class VocabFactory: return vocab, special_vocab -def default_convention_outfile(file_type: GGMLFileType, params: Params, model_params_count: int, metadata: Metadata) -> str: - quantization = { +def default_convention_outfile(file_type: GGMLFileType, expert_count: int | None, model_params_count: tuple[int, int, int], metadata: gguf.Metadata) -> str: + name = metadata.name if metadata.name is not None else None + basename = metadata.basename if metadata.basename is not None else None + finetune = metadata.finetune if metadata.finetune is not None else None + version = metadata.version if metadata.version is not None else None + size_label = metadata.size_label if metadata.size_label is not None else gguf.size_label(*model_params_count, expert_count=expert_count or 0) + + output_type = { GGMLFileType.AllF32: "F32", GGMLFileType.MostlyF16: "F16", GGMLFileType.MostlyQ8_0: "Q8_0", }[file_type] - parameters = model_parameter_count_rounded_notation(model_params_count) - - expert_count = "" - if params.n_experts is not None: - expert_count = f"{params.n_experts}x" - - version = "" - if metadata is not None and metadata.version is not None: - version = f"-{metadata.version}" - - name = "ggml-model" - if metadata is not None and metadata.name is not None: - name = metadata.name - elif params.path_model is not None: - name = params.path_model.name - - return f"{name}{version}-{expert_count}{parameters}-{quantization}" + return gguf.naming_convention(name, basename, finetune, version, size_label, output_type) -def default_outfile(model_paths: list[Path], file_type: GGMLFileType, params: Params, model_params_count: int, metadata: Metadata) -> Path: - default_filename = default_convention_outfile(file_type, params, model_params_count, metadata) +def default_outfile(model_paths: list[Path], file_type: GGMLFileType, expert_count: int | None, model_params_count: tuple[int, int, int], metadata: gguf.Metadata) -> Path: + default_filename = default_convention_outfile(file_type, expert_count, model_params_count, metadata) ret = model_paths[0].parent / f"{default_filename}.gguf" if ret in model_paths: logger.error( @@ -1296,8 +1305,9 @@ def main(args_in: list[str] | None = None) -> None: parser.add_argument("--pad-vocab", action="store_true", help="add pad tokens when model vocab expects more than tokenizer metadata provides") parser.add_argument("--skip-unknown", action="store_true", help="skip unknown tensor names instead of failing") parser.add_argument("--verbose", action="store_true", help="increase output verbosity") - parser.add_argument("--metadata", type=Path, help="Specify the path for a metadata file") + parser.add_argument("--metadata", type=Path, help="Specify the path for an authorship metadata override file") parser.add_argument("--get-outfile", action="store_true", help="get calculated default outfile name") + parser.add_argument("--model-name", type=str, default=None, help="name of the model") args = parser.parse_args(args_in) @@ -1309,32 +1319,36 @@ def main(args_in: list[str] | None = None) -> None: else: logging.basicConfig(level=logging.INFO) - metadata = Metadata.load(args.metadata) + model_name = args.model_name + dir_model = args.model + + metadata = gguf.Metadata.load(args.metadata, dir_model, model_name) if args.get_outfile: - model_plus = load_some_model(args.model) + model_plus = load_some_model(dir_model) params = Params.load(model_plus) - model = convert_model_names(model_plus.model, params, args.skip_unknown) - model_params_count = model_parameter_count(model_plus.model) - ftype = pick_output_type(model, args.outtype) - print(f"{default_convention_outfile(ftype, params, model_params_count, metadata)}") # noqa: NP100 + model = convert_model_names(model_plus.model, params, args.skip_unknown) + model_params_count = per_model_weight_count_estimation(model_plus.model.items()) + ftype = pick_output_type(model, args.outtype) + + if (metadata is None or metadata.name is None) and params.path_model is not None: + metadata.name = params.path_model.name + + print(f"{default_convention_outfile(ftype, params.n_experts, model_params_count, metadata)}") # noqa: NP100 return if args.no_vocab and args.vocab_only: raise ValueError("--vocab-only does not make sense with --no-vocab") if args.dump_single: - model_plus = lazy_load_file(args.model) + model_plus = lazy_load_file(dir_model) do_dump_model(model_plus) return if not args.vocab_only: - model_plus = load_some_model(args.model) + model_plus = load_some_model(dir_model) else: - model_plus = ModelPlus(model = {}, paths = [args.model / 'dummy'], format = 'none', vocab = None) - - model_params_count = model_parameter_count(model_plus.model) - logger.info(f"model parameters count : {model_params_count} ({model_parameter_count_rounded_notation(model_params_count)})") + model_plus = ModelPlus(model = {}, paths = [dir_model / 'dummy'], format = 'none', vocab = None) if args.dump: do_dump_model(model_plus) @@ -1367,7 +1381,7 @@ def main(args_in: list[str] | None = None) -> None: logger.info(f"params = {params}") model_parent_path = model_plus.paths[0].parent - vocab_path = Path(args.vocab_dir or args.model or model_parent_path) + vocab_path = Path(args.vocab_dir or dir_model or model_parent_path) vocab_factory = VocabFactory(vocab_path) vocab_types = None if args.no_vocab else args.vocab_type.split(",") vocab, special_vocab = vocab_factory.load_vocab(vocab_types, model_parent_path) @@ -1396,13 +1410,23 @@ def main(args_in: list[str] | None = None) -> None: if model_plus.vocab is not None and args.vocab_dir is None and not args.no_vocab: vocab = model_plus.vocab + assert params is not None + + if metadata.name is None and params.path_model is not None: + metadata.name = params.path_model.name + + model_params_count = per_model_weight_count_estimation(model_plus.model.items()) + logger.info(f"model parameters count : {model_params_count} ({gguf.model_weight_count_rounded_notation(model_params_count[0])})") + logger.info(f"Vocab info: {vocab}") logger.info(f"Special vocab info: {special_vocab}") model = model_plus.model model = convert_model_names(model, params, args.skip_unknown) ftype = pick_output_type(model, args.outtype) model = convert_to_output_type(model, ftype) - outfile = args.outfile or default_outfile(model_plus.paths, ftype, params, model_params_count, metadata) + outfile = args.outfile or default_outfile(model_plus.paths, ftype, params.n_experts, model_params_count, metadata=metadata) + + metadata.size_label = gguf.size_label(*model_params_count, expert_count=params.n_experts or 0) params.ftype = ftype logger.info(f"Writing {outfile}, format {ftype}") diff --git a/examples/deprecation-warning/README.md b/examples/deprecation-warning/README.md new file mode 100644 index 000000000..59918ec2b --- /dev/null +++ b/examples/deprecation-warning/README.md @@ -0,0 +1,49 @@ +# Migration notice for binary filenames + +> [!IMPORTANT] +[2024 Jun 12] Binaries have been renamed w/ a `llama-` prefix. `main` is now `llama-cli`, `server` is `llama-server`, etc (https://github.com/ggerganov/llama.cpp/pull/7809) + +This migration was important, but it is a breaking change that may not always be immediately obvious to users. + +Please update all scripts and workflows to use the new binary names. + +| Old Filename | New Filename | +| ---- | ---- | +| main | llama-cli | +| server | llama-server | +| llama-bench | llama-bench | +| embedding | llama-embedding | +| quantize | llama-quantize | +| tokenize | llama-tokenize | +| export-lora | llama-export-lora | +| libllava.a | libllava.a | +| baby-llama | llama-baby-llama | +| batched | llama-batched | +| batched-bench | llama-batched-bench | +| benchmark-matmult | llama-benchmark-matmult | +| convert-llama2c-to-ggml | llama-convert-llama2c-to-ggml | +| eval-callback | llama-eval-callback | +| gbnf-validator | llama-gbnf-validator | +| gguf | llama-gguf | +| gguf-split | llama-gguf-split | +| gritlm | llama-gritlm | +| imatrix | llama-imatrix | +| infill | llama-infill | +| llava-cli | llama-llava-cli | +| lookahead | llama-lookahead | +| lookup | llama-lookup | +| lookup-create | llama-lookup-create | +| lookup-merge | llama-lookup-merge | +| lookup-stats | llama-lookup-stats | +| parallel | llama-parallel | +| passkey | llama-passkey | +| perplexity | llama-perplexity | +| q8dot | llama-q8dot | +| quantize-stats | llama-quantize-stats | +| retrieval | llama-retrieval | +| save-load-state | llama-save-load-state | +| simple | llama-simple | +| speculative | llama-speculative | +| vdot | llama-vdot | +| tests/test-c.o | tests/test-c.o | + diff --git a/examples/deprecation-warning/deprecation-warning.cpp b/examples/deprecation-warning/deprecation-warning.cpp new file mode 100644 index 000000000..11b35d2c2 --- /dev/null +++ b/examples/deprecation-warning/deprecation-warning.cpp @@ -0,0 +1,35 @@ +// Warns users that this filename was deprecated, and provides a link for more information. + +#include +#include +#include + +// Main +int main(int argc, char** argv) { + std::string filename = "main"; + if (argc >= 1) { + filename = argv[0]; + } + + // Get only the program name from the full path + auto pos = filename.find_last_of('/'); + if (pos != std::string::npos) { + filename = filename.substr(pos+1); + } + + // Append "llama-" to the beginning of filename to get the replacemnt filename + auto replacement_filename = "llama-" + filename; + + // The exception is if the filename is "main", then our replacement filename is "llama-cli" + if (filename == "main") { + replacement_filename = "llama-cli"; + } + + fprintf(stdout, "\n"); + fprintf(stdout, "WARNING: The binary '%s' is deprecated.\n", filename.c_str()); + fprintf(stdout, " Please use '%s' instead.\n", replacement_filename.c_str()); + fprintf(stdout, " See https://github.com/ggerganov/llama.cpp/tree/master/examples/deprecation-warning/README.md for more information.\n"); + fprintf(stdout, "\n"); + + return EXIT_FAILURE; +} diff --git a/examples/embedding/README.md b/examples/embedding/README.md index 86df18958..e3705b454 100644 --- a/examples/embedding/README.md +++ b/examples/embedding/README.md @@ -58,4 +58,3 @@ The above command will output space-separated float values. ```powershell embedding.exe -p 'Castle<#sep#>Stronghold<#sep#>Dog<#sep#>Cat' --embd-separator '<#sep#>' --embd-normalize 2 --embd-output-format '' -m './path/to/model.gguf' --n-gpu-layers 99 --log-disable 2>/dev/null ``` - diff --git a/examples/eval-callback/eval-callback.cpp b/examples/eval-callback/eval-callback.cpp index 64cd338c2..37d30ab8c 100644 --- a/examples/eval-callback/eval-callback.cpp +++ b/examples/eval-callback/eval-callback.cpp @@ -62,7 +62,7 @@ static void ggml_print_tensor(uint8_t * data, ggml_type type, const int64_t * ne } else if (type == GGML_TYPE_I8) { v = (float) *(int8_t *) &data[i]; } else { - GGML_ASSERT(false); + GGML_ABORT("fatal error"); } printf("%12.4f", v); sum += v; @@ -99,7 +99,7 @@ static bool ggml_debug(struct ggml_tensor * t, bool ask, void * user_data) { char src1_str[128] = {0}; if (src1) { - sprintf(src1_str, "%s{%s}", src1->name, ggml_ne_string(src1).c_str()); + snprintf(src1_str, sizeof(src1_str), "%s{%s}", src1->name, ggml_ne_string(src1).c_str()); } printf("%s: %24s = (%s) %10s(%s{%s}, %s}) = {%s}\n", __func__, diff --git a/examples/export-lora/README.md b/examples/export-lora/README.md index 1fb17feec..91c33c34a 100644 --- a/examples/export-lora/README.md +++ b/examples/export-lora/README.md @@ -6,12 +6,11 @@ Apply LORA adapters to base model and export the resulting model. usage: llama-export-lora [options] options: - -h, --help show this help message and exit - -m FNAME, --model-base FNAME model path from which to load base model (default '') - -o FNAME, --model-out FNAME path to save exported model (default '') - -l FNAME, --lora FNAME apply LoRA adapter - -s FNAME S, --lora-scaled FNAME S apply LoRA adapter with user defined scaling S - -t N, --threads N number of threads to use during computation (default: 4) + -m, --model model path from which to load base model (default '') + --lora FNAME path to LoRA adapter (can be repeated to use multiple adapters) + --lora-scaled FNAME S path to LoRA adapter with user defined scaling S (can be repeated to use multiple adapters) + -t, --threads N number of threads to use during computation (default: 4) + -o, --output FNAME output file (default: 'ggml-lora-merged-f16.gguf') ``` For example: @@ -20,7 +19,15 @@ For example: ./bin/llama-export-lora \ -m open-llama-3b-v2-q8_0.gguf \ -o open-llama-3b-v2-q8_0-english2tokipona-chat.gguf \ - -l lora-open-llama-3b-v2-q8_0-english2tokipona-chat-LATEST.bin + --lora lora-open-llama-3b-v2-q8_0-english2tokipona-chat-LATEST.gguf ``` -Multiple LORA adapters can be applied by passing multiple `-l FN` or `-s FN S` command line parameters. +Multiple LORA adapters can be applied by passing multiple `--lora FNAME` or `--lora-scaled FNAME S` command line parameters: + +```bash +./bin/llama-export-lora \ + -m your_base_model.gguf \ + -o your_merged_model.gguf \ + --lora-scaled lora_task_A.gguf 0.5 \ + --lora-scaled lora_task_B.gguf 0.5 +``` diff --git a/examples/export-lora/export-lora.cpp b/examples/export-lora/export-lora.cpp index 08413f57e..150f7e8d5 100644 --- a/examples/export-lora/export-lora.cpp +++ b/examples/export-lora/export-lora.cpp @@ -1,462 +1,420 @@ - #include "common.h" #include "ggml.h" #include "ggml-alloc.h" +#include #include #include #include +#include -struct lora_info { - std::string filename; +static bool g_verbose = false; + +static std::string get_kv_str(struct gguf_context * ctx_gguf, const std::string & key){ + int id = gguf_find_key(ctx_gguf, key.c_str()); + return id < 0 ? "" : std::string(gguf_get_val_str(ctx_gguf, id)); +} + +static float get_kv_f32(struct gguf_context * ctx_gguf, const std::string & key) { + int id = gguf_find_key(ctx_gguf, key.c_str()); + return id < 0 ? 0.0f : gguf_get_val_f32(ctx_gguf, id); +} + +static void zeros(std::ofstream & file, size_t n) { + char zero = 0; + for (size_t i = 0; i < n; ++i) { + file.write(&zero, 1); + } +} + +static std::string ggml_ne_string(const ggml_tensor * t) { + std::string str; + for (int i = 0; i < GGML_MAX_DIMS; ++i) { + str += std::to_string(t->ne[i]); + if (i + 1 < GGML_MAX_DIMS) { + str += ", "; + } + } + return str; +} + +static struct gguf_context * load_gguf(std::string & fname, struct ggml_context ** ctx_ggml) { + struct gguf_init_params params = { + /*.no_alloc = */ true, + /*.ctx = */ ctx_ggml, + }; + struct gguf_context * ctx_gguf = gguf_init_from_file(fname.c_str(), params); + if (!ctx_gguf) { + throw std::runtime_error("failed to load input GGUF from " + fname); + } + return ctx_gguf; +} + +static void replace_all(std::string & s, const std::string & search, const std::string & replace) { + std::string result; + for (size_t pos = 0; ; pos += search.length()) { + auto new_pos = s.find(search, pos); + if (new_pos == std::string::npos) { + result += s.substr(pos, s.size() - pos); + break; + } + result += s.substr(pos, new_pos - pos) + replace; + pos = new_pos; + } + s = std::move(result); +} + +struct file_input { + struct ggml_context * ctx_meta = nullptr; + struct gguf_context * ctx_gguf = nullptr; + std::ifstream f_in; + std::map tensors; + float alpha; float scale; + + file_input(std::string & fname, float scale): f_in(fname, std::ios::binary), scale(scale) { + if (!f_in.is_open()) { + throw std::runtime_error("failed to open input gguf from " + fname); + } + + ctx_gguf = load_gguf(fname, &ctx_meta); + alpha = get_kv_f32(ctx_gguf, "adapter.lora.alpha"); + printf("%s: loaded gguf from %s\n", __func__, fname.c_str()); + + for (ggml_tensor * cur = ggml_get_first_tensor(ctx_meta); cur; cur = ggml_get_next_tensor(ctx_meta, cur)) { + std::string name(cur->name); + tensors[name] = cur; + if (g_verbose) { + printf("%s: %s\n", __func__, cur->name); + } + } + } + + ggml_tensor * get_tensor(std::string name) { + if (tensors.find(name) == tensors.end()) { + return nullptr; + } + return tensors[name]; + } + + void read_tensor_data(std::string name, std::vector & buf) { + if (tensors.find(name) == tensors.end()) { + throw std::runtime_error("cannot find tensor with name: " + name); + } + auto len = ggml_nbytes(tensors[name]); + if (buf.size() < len) { + buf.resize(len); + } + auto i_tensor_in = gguf_find_tensor(ctx_gguf, name.c_str()); // idx of tensor in the input file + auto offset = gguf_get_data_offset(ctx_gguf) + gguf_get_tensor_offset(ctx_gguf, i_tensor_in); + f_in.seekg(offset); + f_in.read((char* )buf.data(), len); + } + + ~file_input() { + gguf_free(ctx_gguf); + ggml_free(ctx_meta); + } }; -struct export_lora_params { - std::string fn_model_base; - std::string fn_model_out; - std::vector lora; +struct lora_merge_ctx { + // input base model + adapters + file_input base_model; + std::vector> adapters; + + // for computing merged tensor int n_threads; -}; + ggml_backend_t backend = nullptr; + ggml_gallocr_t allocr = nullptr; + std::vector read_buf; -struct lora_data { - struct lora_info info; - std::vector data; - struct ggml_context * ctx; + // output file + struct gguf_context * ctx_out; + struct ggml_context * ctx_out_ggml; + std::ofstream fout; - uint32_t lora_r; - uint32_t lora_alpha; -}; + lora_merge_ctx( + std::string & base_fname, + std::vector> & lora_files, + std::string & outfile, + int n_threads) : base_model(base_fname, 0), n_threads(n_threads), fout(outfile, std::ios::binary) { + fout.exceptions(std::ofstream::failbit); // fail fast on write errors -struct llama_file { - // use FILE * so we don't have to re-open the file to mmap - FILE * fp; - size_t size; + if (gguf_find_key(base_model.ctx_gguf, LLM_KV_SPLIT_COUNT) >= 0) { + throw std::runtime_error("split model is not yet supported"); + } - llama_file(const char * fname, const char * mode) { - fp = std::fopen(fname, mode); - if (fp == NULL) { - size = 0; + for (auto lora_inp : lora_files) { + auto fname = std::get<0>(lora_inp); + auto scale = std::get<1>(lora_inp); + std::unique_ptr adapter(new file_input(fname, scale)); + check_metadata_lora(adapter.get()); + adapters.push_back(std::move(adapter)); + } + + ctx_out = gguf_init_empty(); + struct ggml_init_params params = { + /*.mem_size =*/ gguf_get_n_tensors(base_model.ctx_gguf)*ggml_tensor_overhead(), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + ctx_out_ggml = ggml_init(params); + backend = ggml_backend_cpu_init(); + allocr = ggml_gallocr_new(ggml_backend_get_default_buffer_type(backend)); + } + + void check_metadata_lora(file_input * adapter) { + auto general_type = get_kv_str(adapter->ctx_gguf, "general.type"); + if (general_type != "adapter") { + throw std::runtime_error("expect general.type to be 'adapter', but got: " + general_type); + } + + auto adapter_type = get_kv_str(adapter->ctx_gguf, "adapter.type"); + if (adapter_type != "lora") { + throw std::runtime_error("expect adapter.type to be 'lora', but got: " + adapter_type); + } + + auto general_arch_base = get_kv_str(base_model.ctx_gguf, "general.architecture"); + auto general_arch_lora = get_kv_str(adapter->ctx_gguf, "general.architecture"); + if (general_arch_base != general_arch_lora) { + throw std::runtime_error("model arch and LoRA arch mismatch"); + } + } + + ggml_type get_out_tensor_type(struct ggml_tensor * t) { + if (t->type == GGML_TYPE_F32) { + return GGML_TYPE_F32; } else { - seek(0, SEEK_END); - size = tell(); - seek(0, SEEK_SET); + return GGML_TYPE_F16; } } - size_t tell() const { -#ifdef _WIN32 - __int64 ret = _ftelli64(fp); -#else - long ret = std::ftell(fp); -#endif - GGML_ASSERT(ret != -1); // this really shouldn't fail - return (size_t) ret; - } + void run_merge() { + // prepare metadata + gguf_set_kv(ctx_out, base_model.ctx_gguf); + // output is forced to f16 for now + gguf_set_val_u32(ctx_out, "general.file_type", LLAMA_FTYPE_MOSTLY_F16); - void seek(size_t offset, int whence) { -#ifdef _WIN32 - int ret = _fseeki64(fp, (__int64) offset, whence); -#else - int ret = std::fseek(fp, (long) offset, whence); -#endif - GGML_ASSERT(ret == 0); // same - } - - void read_raw(void * ptr, size_t size) { - if (size == 0) { - return; + // check if all lora adapters have the same tensors + // TODO: remove this when we can support merging subset of adapters. Ref: https://github.com/ggerganov/llama.cpp/pull/8607#discussion_r1686027777 + static const char * err_no_subset_adapter = "Input adapters do not have the same list of tensors. This is not yet supported. Please merge the adapter one-by-one instead of merging all at once."; + if (adapters.size() > 1) { + for (size_t i = 1; i < adapters.size(); ++i) { + if (adapters[0]->tensors.size() != adapters[i]->tensors.size()) { + throw std::runtime_error(err_no_subset_adapter); + } + for (auto & it : adapters[i]->tensors) { + if (adapters[0]->get_tensor(it.first) == nullptr) { + throw std::runtime_error(err_no_subset_adapter); + } + } + } } - errno = 0; - std::size_t ret = std::fread(ptr, size, 1, fp); - if (ferror(fp)) { - die_fmt("read error: %s", strerror(errno)); + + // mapping base tensor to out tensor (same shape with base, but different type) + // if out_tensor == nullptr, we only copy it + std::vector> base_to_out_tensors; + for (auto & it : base_model.tensors) { + bool t_a = true; + bool t_b = true; + for (auto & adapter : adapters) { + t_a &= nullptr != adapter->get_tensor(it.first + ".lora_a"); + t_b &= nullptr != adapter->get_tensor(it.first + ".lora_b"); + } + auto base_tensor = it.second; + if (!t_a && !t_b) { + // only copy + struct ggml_tensor * cpy_tensor = ggml_dup_tensor(ctx_out_ggml, base_tensor); + ggml_set_name(cpy_tensor, base_tensor->name); + base_to_out_tensors.push_back(std::make_pair(cpy_tensor, nullptr)); + gguf_add_tensor(ctx_out, cpy_tensor); + } else if (t_a && t_b) { + // need merging + struct ggml_tensor * out_tensor = ggml_new_tensor( + ctx_out_ggml, get_out_tensor_type(base_tensor), GGML_MAX_DIMS, base_tensor->ne); + ggml_set_name(out_tensor, base_tensor->name); + base_to_out_tensors.push_back(std::make_pair(base_tensor, out_tensor)); + gguf_add_tensor(ctx_out, out_tensor); + } else { + throw std::runtime_error("tensor " + it.first + " missing either lora_a or lora_b"); + } } - if (ret != 1) { - die("unexpectedly reached end of file"); + + // placeholder for the meta data + { + size_t meta_size = gguf_get_meta_size(ctx_out); + zeros(fout, meta_size); } - } - std::uint32_t read_u32() { - std::uint32_t ret; - read_raw(&ret, sizeof(ret)); - return ret; - } - - std::string read_string(std::uint32_t len) { - std::vector chars(len); - read_raw(chars.data(), len); - return std::string(chars.data(), len); - } - - void write_raw(const void * ptr, size_t size) { - if (size == 0) { - return; + // process base model tensors + size_t n_merged = 0; + for (auto & it : base_to_out_tensors) { + if (it.second != nullptr) { + merge_tensor(it.first, it.second); + n_merged++; + } else { + copy_tensor(it.first); + } } - errno = 0; - size_t ret = std::fwrite(ptr, size, 1, fp); - if (ret != 1) { - die_fmt("write error: %s", strerror(errno)); + + // write output metadata + { + std::vector data(gguf_get_meta_size(ctx_out)); + gguf_get_meta_data(ctx_out, data.data()); + fout.seekp(0); + fout.write((const char *)data.data(), data.size()); } + + printf("%s : merged %ld tensors with lora adapters\n", __func__, n_merged); + printf("%s : wrote %ld tensors to output file\n", __func__, base_to_out_tensors.size()); } - void write_u32(std::uint32_t val) { - write_raw(&val, sizeof(val)); + void copy_tensor(struct ggml_tensor * base) { + printf("%s : %s [%s]\n", __func__, base->name, ggml_ne_string(base).c_str()); + size_t len = ggml_nbytes(base); + base_model.read_tensor_data(base->name, read_buf); + fout.write((char* )read_buf.data(), len); + zeros(fout, GGML_PAD(len, GGUF_DEFAULT_ALIGNMENT) - len); } - bool eof() { - return tell() >= size; - } + void merge_tensor(struct ggml_tensor * base, struct ggml_tensor * out) { + std::string name_base(base->name); + std::string name_lora_a = name_base + ".lora_a"; + std::string name_lora_b = name_base + ".lora_b"; - ~llama_file() { - if (fp) { - std::fclose(fp); + printf("%s : %s [%s]\n", __func__, base->name, ggml_ne_string(base).c_str()); + + // context for input tensor + std::vector inp_a(adapters.size()); + std::vector inp_b(adapters.size()); + struct ggml_init_params params { + /*.mem_size =*/ ggml_tensor_overhead()*(2+adapters.size()*2), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + struct ggml_context * ctx = ggml_init(params); + + // alloc tensors + struct ggml_tensor * inp_base = ggml_new_tensor(ctx, GGML_TYPE_F32, GGML_MAX_DIMS, base->ne); + for (size_t i = 0; i < adapters.size(); ++i) { + auto t_a = adapters[i]->get_tensor(name_lora_a); + auto t_b = adapters[i]->get_tensor(name_lora_b); + inp_a[i] = ggml_dup_tensor(ctx, t_a); + inp_b[i] = ggml_dup_tensor(ctx, t_b); } + ggml_backend_buffer_t buffer = ggml_backend_alloc_ctx_tensors(ctx, backend); + + // load base tensor to backend buffer + base_model.read_tensor_data(name_base, read_buf); + if (base->type != GGML_TYPE_F32) { + // optionally dequantize it + printf("%s : + dequantize base tensor from %s to F32\n", __func__, ggml_type_name(base->type)); + auto nels = ggml_nelements(inp_base); + ggml_type_traits_t qtype = ggml_internal_get_type_traits(base->type); + std::vector dequant_buf(nels * sizeof(float)); + qtype.to_float(read_buf.data(), (float *)dequant_buf.data(), nels); + ggml_backend_tensor_set(inp_base, dequant_buf.data(), 0, dequant_buf.size()); + } else { + ggml_backend_tensor_set(inp_base, read_buf.data(), 0, ggml_nbytes(inp_base)); + } + + // load lora tensors to backend buffer + for (size_t i = 0; i < adapters.size(); ++i) { + adapters[i]->read_tensor_data(name_lora_a, read_buf); + ggml_backend_tensor_set(inp_a[i], read_buf.data(), 0, ggml_nbytes(inp_a[i])); + adapters[i]->read_tensor_data(name_lora_b, read_buf); + ggml_backend_tensor_set(inp_b[i], read_buf.data(), 0, ggml_nbytes(inp_b[i])); + } + + // build graph + struct ggml_cgraph * gf; + { + static size_t buf_size = ggml_tensor_overhead()*GGML_DEFAULT_GRAPH_SIZE + ggml_graph_overhead(); + static std::vector buf(buf_size); + struct ggml_init_params params0 = { + /*.mem_size =*/ buf_size, + /*.mem_buffer =*/ buf.data(), + /*.no_alloc =*/ true, + }; + struct ggml_context * ctx0 = ggml_init(params0); + gf = ggml_new_graph(ctx0); + struct ggml_tensor * cur = inp_base; + for (size_t i = 0; i < adapters.size(); ++i) { + struct ggml_tensor * a_T = ggml_cont(ctx0, ggml_transpose(ctx0, ggml_cast(ctx0, inp_a[i], GGML_TYPE_F32))); + struct ggml_tensor * delta = ggml_mul_mat(ctx0, a_T, ggml_cast(ctx0, inp_b[i], GGML_TYPE_F32)); + // scale + const float alpha = adapters[i]->alpha; + const float rank = (float) inp_b[i]->ne[0]; + const float scale = alpha ? adapters[i]->scale * alpha / rank : adapters[i]->scale; + delta = ggml_scale(ctx0, delta, scale); + cur = ggml_add(ctx0, delta, cur); + printf("%s : + merging from adapter[%ld] type=%s\n", __func__, i, ggml_type_name(inp_a[i]->type)); + printf("%s : input_scale=%f calculated_scale=%f rank=%d\n", __func__, adapters[i]->scale, scale, (int) inp_b[i]->ne[0]); + } + cur = ggml_cast(ctx0, cur, out->type); + printf("%s : + output type is %s\n", __func__, ggml_type_name(out->type)); + ggml_build_forward_expand(gf, cur); + ggml_free(ctx0); + } + + // compute + { + ggml_gallocr_alloc_graph(allocr, gf); + ggml_backend_cpu_set_n_threads(backend, n_threads); + ggml_backend_graph_compute(backend, gf); + } + + // write data to output file + { + auto result = gf->nodes[gf->n_nodes - 1]; + size_t len = ggml_nbytes(result); + if (read_buf.size() < len) { + read_buf.resize(len); + } + ggml_backend_tensor_get(result, read_buf.data(), 0, len); + fout.write((char* )read_buf.data(), len); + zeros(fout, GGML_PAD(len, GGUF_DEFAULT_ALIGNMENT) - len); + } + + ggml_free(ctx); + ggml_backend_buffer_free(buffer); + } + + ~lora_merge_ctx() { + ggml_gallocr_free(allocr); + ggml_backend_free(backend); + gguf_free(ctx_out); + ggml_free(ctx_out_ggml); } }; -static struct export_lora_params get_default_export_lora_params() { - struct export_lora_params result; - result.fn_model_base = ""; - result.fn_model_out = ""; - result.n_threads = GGML_DEFAULT_N_THREADS; - return result; -} +static void print_usage(int argc, char ** argv, const gpt_params & params) { + gpt_params_print_usage(argc, argv, params); -static void export_lora_print_usage(int /*argc*/, char ** argv, const struct export_lora_params * params) { - fprintf(stderr, "usage: %s [options]\n", argv[0]); - fprintf(stderr, "\n"); - fprintf(stderr, "options:\n"); - fprintf(stderr, " -h, --help show this help message and exit\n"); - fprintf(stderr, " -m FNAME, --model-base FNAME model path from which to load base model (default '%s')\n", params->fn_model_base.c_str()); - fprintf(stderr, " -o FNAME, --model-out FNAME path to save exported model (default '%s')\n", params->fn_model_out.c_str()); - fprintf(stderr, " -l FNAME, --lora FNAME apply LoRA adapter\n"); - fprintf(stderr, " -s FNAME S, --lora-scaled FNAME S apply LoRA adapter with user defined scaling S\n"); - fprintf(stderr, " -t N, --threads N number of threads to use during computation (default: %d)\n", params->n_threads); -} - -static bool export_lora_params_parse(int argc, char ** argv, struct export_lora_params * params) { - bool invalid_param = false; - std::string arg; - struct export_lora_params default_params = get_default_export_lora_params(); - const std::string arg_prefix = "--"; - - for (int i = 1; i < argc; i++) { - arg = argv[i]; - if (arg.compare(0, arg_prefix.size(), arg_prefix) == 0) { - std::replace(arg.begin(), arg.end(), '_', '-'); - } - - if (arg == "-m" || arg == "--model-base") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->fn_model_base = argv[i]; - } else if (arg == "-o" || arg == "--model-out") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->fn_model_out = argv[i]; - } else if (arg == "-l" || arg == "--lora") { - if (++i >= argc) { - invalid_param = true; - break; - } - struct lora_info lora; - lora.filename = argv[i]; - lora.scale = 1.0f; - params->lora.push_back(lora); - } else if (arg == "-s" || arg == "--lora-scaled") { - if (++i >= argc) { - invalid_param = true; - break; - } - struct lora_info lora; - lora.filename = argv[i]; - if (++i >= argc) { - invalid_param = true; - break; - } - lora.scale = std::stof(argv[i]); - params->lora.push_back(lora); - } else if (arg == "-t" || arg == "--threads") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_threads = std::stoi(argv[i]); - if (params->n_threads <= 0) { - params->n_threads = std::thread::hardware_concurrency(); - } - } else { - fprintf(stderr, "error: unknown argument: '%s'\n", arg.c_str()); - export_lora_print_usage(argc, argv, &default_params); - exit(1); - } - } - - if (params->fn_model_base == default_params.fn_model_base) { - fprintf(stderr, "error: please specify a filename for model-base.\n"); - export_lora_print_usage(argc, argv, &default_params); - exit(1); - } - if (params->fn_model_out == default_params.fn_model_out) { - fprintf(stderr, "error: please specify a filename for model-out.\n"); - export_lora_print_usage(argc, argv, &default_params); - exit(1); - } - if (invalid_param) { - fprintf(stderr, "error: invalid parameter for argument: '%s'\n", arg.c_str()); - export_lora_print_usage(argc, argv, &default_params); - exit(1); - } - return true; -} - -static void free_lora(struct lora_data * lora) { - if (lora->ctx != NULL) { - ggml_free(lora->ctx); - } - delete lora; -} - -static struct lora_data * load_lora(struct lora_info * info) { - struct lora_data * result = new struct lora_data; - result->info = *info; - result->ctx = NULL; - result->lora_r = 1; - result->lora_alpha = 1; - - struct llama_file file(info->filename.c_str(), "rb"); - if (file.fp == NULL) { - fprintf(stderr, "warning: Could not open lora adapter '%s'. Ignoring this adapter.\n", - info->filename.c_str()); - free_lora(result); - return NULL; - } - - struct ggml_init_params params_ggml; - params_ggml.mem_size = ggml_tensor_overhead() * GGML_DEFAULT_GRAPH_SIZE; - params_ggml.mem_buffer = NULL; - params_ggml.no_alloc = true; - result->ctx = ggml_init(params_ggml); - - uint32_t magic = file.read_u32(); - if (magic != LLAMA_FILE_MAGIC_GGLA) { - die_fmt("unexpected lora header file magic in '%s'", info->filename.c_str()); - } - uint32_t version = file.read_u32(); - if (version != 1) { - die_fmt("unexpected lora file version '%u' in '%s'", (unsigned) version, info->filename.c_str()); - } - result->lora_r = file.read_u32(); - result->lora_alpha = file.read_u32(); - // read tensor infos from file - std::vector name_buf; - std::vector tensors; - std::vector tensors_offset; - size_t total_nbytes_pad = 0; - while(!file.eof()) { - int64_t ne[4] = {1,1,1,1}; - uint32_t n_dims = file.read_u32(); - uint32_t namelen = file.read_u32(); - uint32_t type = file.read_u32(); - for (uint32_t k = 0; k < n_dims; ++k) { - ne[k] = (int64_t)file.read_u32(); - } - name_buf.clear(); - name_buf.resize(namelen + 1, '\0'); - file.read_raw(name_buf.data(), namelen); - file.seek((0-file.tell()) & 31, SEEK_CUR); - size_t offset = file.tell(); - struct ggml_tensor * tensor = ggml_new_tensor(result->ctx, (enum ggml_type) type, n_dims, ne); - ggml_set_name(tensor, name_buf.data()); - size_t nbytes = ggml_nbytes(tensor); - size_t nbytes_pad = ggml_nbytes_pad(tensor); - total_nbytes_pad += nbytes_pad; - tensors.push_back(tensor); - tensors_offset.push_back(offset); - file.seek(nbytes, SEEK_CUR); - } - // read tensor data - result->data.resize(total_nbytes_pad); - size_t data_offset = 0; - for (size_t i = 0; i < tensors.size(); ++i) { - struct ggml_tensor * tensor = tensors[i]; - size_t offset = tensors_offset[i]; - size_t nbytes = ggml_nbytes(tensor); - size_t nbytes_pad = ggml_nbytes_pad(tensor); - file.seek(offset, SEEK_SET); - tensor->data = result->data.data() + data_offset; - file.read_raw(tensor->data, nbytes); - data_offset += nbytes_pad; - } - return result; -} - - -static struct ggml_cgraph * build_graph_lora( - struct ggml_context * ctx, - struct ggml_tensor * tensor, - struct ggml_tensor * lora_a, - struct ggml_tensor * lora_b, - float scaling -) { - struct ggml_tensor * ab = ggml_mul_mat(ctx, lora_a, lora_b); - if (scaling != 1.0f) { - ab = ggml_scale(ctx, ab, scaling); - } - struct ggml_tensor * res = ggml_add_inplace(ctx, tensor, ab); - - struct ggml_cgraph * gf = ggml_new_graph(ctx); - ggml_build_forward_expand (gf, res); - return gf; -} - -static bool apply_lora(struct ggml_tensor * tensor, struct lora_data * lora, int n_threads) { - if (lora->ctx == NULL) { - return false; - } - std::string name = ggml_get_name(tensor); - std::string name_a = name + std::string(".loraA"); - std::string name_b = name + std::string(".loraB"); - struct ggml_tensor * lora_a = ggml_get_tensor(lora->ctx, name_a.c_str()); - struct ggml_tensor * lora_b = ggml_get_tensor(lora->ctx, name_b.c_str()); - if (lora_a == NULL || lora_b == NULL) { - return false; - } - - float scaling = lora->info.scale * (float)lora->lora_alpha / (float)lora->lora_r; - - struct ggml_init_params params; - params.mem_size = GGML_OBJECT_SIZE + ggml_graph_overhead() + ggml_tensor_overhead()*4 + GGML_MEM_ALIGN*5; - params.mem_buffer = NULL; - params.no_alloc = true; - struct ggml_context * ctx = NULL; - struct ggml_gallocr * alloc = NULL; - struct ggml_cgraph * gf = NULL; - - ctx = ggml_init(params); - alloc = ggml_gallocr_new(ggml_backend_cpu_buffer_type()); - gf = build_graph_lora(ctx, tensor, lora_a, lora_b, scaling); - - ggml_gallocr_alloc_graph(alloc, gf); - - struct ggml_cplan cplan = ggml_graph_plan(gf, n_threads); - static std::vector data_work; - data_work.resize(cplan.work_size); - cplan.work_data = data_work.data(); - - ggml_graph_compute(gf, &cplan); - - ggml_gallocr_free(alloc); - ggml_free(ctx); - return true; -} - -static void export_lora(struct export_lora_params * params) { - // load all loras - std::vector loras; - for (size_t i = 0; i < params->lora.size(); ++i) { - struct lora_data * lora = load_lora(¶ms->lora[i]); - if (lora != NULL) { - loras.push_back(lora); - } - } - if (loras.size() == 0) { - fprintf(stderr, "warning: no lora adapters will be applied.\n"); - } - - // open input file - struct llama_file fin(params->fn_model_base.c_str(), "rb"); - if (!fin.fp) { - die_fmt("Could not open file '%s'\n", params->fn_model_base.c_str()); - } - - // open base model gguf, read tensors without their data - struct ggml_context * ctx_in; - struct gguf_init_params params_gguf; - params_gguf.no_alloc = true; - params_gguf.ctx = &ctx_in; - struct gguf_context * gguf_in = gguf_init_from_file(params->fn_model_base.c_str(), params_gguf); - - // create new gguf - struct gguf_context * gguf_out = gguf_init_empty(); - - // copy meta data from base model: kv and tensors - gguf_set_kv(gguf_out, gguf_in); - int n_tensors = gguf_get_n_tensors(gguf_in); - for (int i=0; i < n_tensors; ++i) { - const char * name = gguf_get_tensor_name(gguf_in, i); - struct ggml_tensor * tensor = ggml_get_tensor(ctx_in, name); - gguf_add_tensor(gguf_out, tensor); - } - - // create output file - struct llama_file fout(params->fn_model_out.c_str(), "wb"); - if (!fout.fp) { - die_fmt("Could not create file '%s'\n", params->fn_model_out.c_str()); - } - - // write gguf meta data - std::vector meta; - meta.resize(gguf_get_meta_size(gguf_out)); - gguf_get_meta_data(gguf_out, meta.data()); - fout.write_raw(meta.data(), meta.size()); - - std::vector data; - std::vector padding; - for (int i=0; i < n_tensors; ++i) { - const char * name = gguf_get_tensor_name(gguf_in, i); - struct ggml_tensor * tensor = ggml_get_tensor(ctx_in, name); - - // read tensor data - data.resize(ggml_nbytes(tensor)); - tensor->data = data.data(); - size_t offset = gguf_get_tensor_offset(gguf_in, i); - fin.seek(offset + meta.size(), SEEK_SET); - fin.read_raw(data.data(), data.size()); - - // apply all loras - for (size_t k = 0; k < loras.size(); ++k) { - apply_lora(tensor, loras[k], params->n_threads); - } - - // write tensor data + padding - padding.clear(); - padding.resize(GGML_PAD(data.size(), gguf_get_alignment(gguf_out)) - data.size(), 0); - - GGML_ASSERT(fout.tell() == offset + meta.size()); - // fout.seek(offset + meta.size(), SEEK_SET); - fout.write_raw(data.data(), data.size()); - fout.write_raw(padding.data(), padding.size()); - - if (i % 2 == 0) { - printf("."); - } - } + printf("\nexample usage:\n"); + printf("\n %s -m base-model.gguf --lora lora-file.gguf -o merged-model-f16.gguf\n", argv[0]); + printf("\nNOTE: output model is F16\n"); printf("\n"); - - // close gguf - gguf_free(gguf_out); - gguf_free(gguf_in); - - // free loras - for (size_t i = 0; i < loras.size(); ++i) { - free_lora(loras[i]); - } } int main(int argc, char ** argv) { - struct export_lora_params params = get_default_export_lora_params(); + gpt_params params; - if (!export_lora_params_parse(argc, argv, ¶ms)) { + if (!gpt_params_parse(argc, argv, params)) { + print_usage(argc, argv, params); return 1; } - export_lora(¶ms); + g_verbose = (params.verbosity == 1); + try { + lora_merge_ctx ctx(params.model, params.lora_adapter, params.lora_outfile, params.n_threads); + ctx.run_merge(); + } catch (const std::exception & err) { + fprintf(stderr, "%s\n", err.what()); + exit(EXIT_FAILURE); + } + + printf("done, output file is %s\n", params.lora_outfile.c_str()); return 0; } diff --git a/examples/finetune/CMakeLists.txt b/examples/finetune/CMakeLists.txt deleted file mode 100644 index 64afe6ddc..000000000 --- a/examples/finetune/CMakeLists.txt +++ /dev/null @@ -1,5 +0,0 @@ -set(TARGET llama-finetune) -add_executable(${TARGET} finetune.cpp) -install(TARGETS ${TARGET} RUNTIME) -target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) -target_compile_features(${TARGET} PRIVATE cxx_std_11) diff --git a/examples/finetune/README.md b/examples/finetune/README.md deleted file mode 100644 index a6ae64983..000000000 --- a/examples/finetune/README.md +++ /dev/null @@ -1,90 +0,0 @@ -# finetune - -Basic usage instructions: - -```bash -# get training data -wget https://raw.githubusercontent.com/brunoklein99/deep-learning-notes/master/shakespeare.txt - -# finetune LORA adapter -./bin/llama-finetune \ - --model-base open-llama-3b-v2-q8_0.gguf \ - --checkpoint-in chk-lora-open-llama-3b-v2-q8_0-shakespeare-LATEST.gguf \ - --checkpoint-out chk-lora-open-llama-3b-v2-q8_0-shakespeare-ITERATION.gguf \ - --lora-out lora-open-llama-3b-v2-q8_0-shakespeare-ITERATION.bin \ - --train-data "shakespeare.txt" \ - --save-every 10 \ - --threads 6 --adam-iter 30 --batch 4 --ctx 64 \ - --use-checkpointing - -# predict -./bin/llama-cli -m open-llama-3b-v2-q8_0.gguf --lora lora-open-llama-3b-v2-q8_0-shakespeare-LATEST.bin -``` - -**Only llama based models are supported!** The output files will be saved every N iterations (config with `--save-every N`). -The pattern 'ITERATION' in the output filenames will be replaced with the iteration number and with 'LATEST' for the latest output. -So in above example after 10 iterations these files will be written: -- chk-lora-open-llama-3b-v2-q8_0-shakespeare-10.gguf -- chk-lora-open-llama-3b-v2-q8_0-shakespeare-LATEST.gguf -- lora-open-llama-3b-v2-q8_0-shakespeare-10.bin -- lora-open-llama-3b-v2-q8_0-shakespeare-LATEST.bin - -After 10 more iterations: -- chk-lora-open-llama-3b-v2-q8_0-shakespeare-20.gguf -- chk-lora-open-llama-3b-v2-q8_0-shakespeare-LATEST.gguf -- lora-open-llama-3b-v2-q8_0-shakespeare-20.bin -- lora-open-llama-3b-v2-q8_0-shakespeare-LATEST.bin - -Checkpoint files (`--checkpoint-in FN`, `--checkpoint-out FN`) store the training process. When the input checkpoint file does not exist, it will begin finetuning a new randomly initialized adapter. - -llama.cpp compatible LORA adapters will be saved with filename specified by `--lora-out FN`. -These LORA adapters can then be used by `llama-cli` together with the base model, like in the 'predict' example command above. - -In `llama-cli` you can also load multiple LORA adapters, which will then be mixed together. - -For example if you have two LORA adapters `lora-open-llama-3b-v2-q8_0-shakespeare-LATEST.bin` and `lora-open-llama-3b-v2-q8_0-bible-LATEST.bin`, you can mix them together like this: - -```bash -./bin/llama-cli -m open-llama-3b-v2-q8_0.gguf \ - --lora lora-open-llama-3b-v2-q8_0-shakespeare-LATEST.bin \ - --lora lora-open-llama-3b-v2-q8_0-bible-LATEST.bin -``` - -You can change how strong each LORA adapter is applied to the base model by using `--lora-scaled FN SCALE` instead of `--lora FN`. - -For example to apply 40% of the 'shakespeare' LORA adapter, 80% of the 'bible' LORA adapter and 100% of yet another one: - -```bash -./bin/llama-cli -m open-llama-3b-v2-q8_0.gguf \ - --lora-scaled lora-open-llama-3b-v2-q8_0-shakespeare-LATEST.bin 0.4 \ - --lora-scaled lora-open-llama-3b-v2-q8_0-bible-LATEST.bin 0.8 \ - --lora lora-open-llama-3b-v2-q8_0-yet-another-one-LATEST.bin -``` - -The scale numbers don't need to add up to one, and you can also use numbers greater than 1 to further increase the influence of an adapter. But making the values too big will sometimes result in worse output. Play around to find good values. - -Gradient checkpointing reduces the memory requirements by ~50% but increases the runtime. -If you have enough RAM, you can make finetuning a bit faster by disabling checkpointing with `--no-checkpointing`. - -The default LORA rank can be specified with `--lora-r N`. -The LORA rank can be configured for each model tensor type separately with these command line options: - -```bash - --lora-r N LORA r: default rank. Also specifies resulting scaling together with lora-alpha. (default 4) - --rank-att-norm N LORA rank for attention norm tensor (default 1) - --rank-ffn-norm N LORA rank for feed-forward norm tensor (default 1) - --rank-out-norm N LORA rank for output norm tensor (default 1) - --rank-tok-embd N LORA rank for token embeddings tensor (default 4) - --rank-out N LORA rank for output tensor (default 4) - --rank-wq N LORA rank for wq tensor (default 4) - --rank-wk N LORA rank for wk tensor (default 4) - --rank-wv N LORA rank for wv tensor (default 4) - --rank-wo N LORA rank for wo tensor (default 4) - --rank-ffn_gate N LORA rank for ffn_gate tensor (default 4) - --rank-ffn_down N LORA rank for ffn_down tensor (default 4) - --rank-ffn_up N LORA rank for ffn_up tensor (default 4) -``` - -The LORA rank of 'norm' tensors should always be 1. - -To see all available options use `finetune --help`. diff --git a/examples/finetune/convert-finetune-checkpoint-to-gguf.py b/examples/finetune/convert-finetune-checkpoint-to-gguf.py deleted file mode 100644 index c89090918..000000000 --- a/examples/finetune/convert-finetune-checkpoint-to-gguf.py +++ /dev/null @@ -1,487 +0,0 @@ -#!/usr/bin/env python3 -# finetune checkpoint --> gguf conversion - -import argparse -import gguf -import struct -import numpy as np -from pathlib import Path - -# gguf constants -LLM_KV_OPTIMIZER_TYPE = "optimizer.type" -LLM_KV_OPTIMIZER_TYPE_ADAM = "adam" -LLM_KV_OPTIMIZER_TYPE_LBFGS = "lbfgs" -LLM_KV_OPTIMIZER_FILE_VERSION = "optimizer.file_version" -LLM_KV_OPTIMIZER_CONVERGENCE_PAST_COUNT = "optimizer.convergence_past_count" -LLM_KV_OPTIMIZER_PARAMETER_COUNT = "optimizer.parameter_count" -LLM_KV_OPTIMIZER_ITERATION_COUNT = "optimizer.iteration_count" -LLM_KV_OPTIMIZER_JUST_INITIALIZED = "optimizer.just_initialized" -LLM_KV_OPTIMIZER_ADAM_BEST_LOSS = "optimizer.adam.best_loss" -LLM_KV_OPTIMIZER_ADAM_PREVIOUS_LOSS = "optimizer.adam.previous_loss" -LLM_KV_OPTIMIZER_ADAM_NO_IMPROVEMENT_COUNT = "optimizer.adam.no_improvement_count" -LLM_KV_OPTIMIZER_LBFGS_APPROX_HESSIAN_COUNT = "optimizer.lbfgs.approx_hessian_count" -LLM_KV_OPTIMIZER_LBFGS_BEST_LOSS = "optimizer.lbfgs.best_loss" -LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_STEP = "optimizer.lbfgs.line_search_step" -LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_J = "optimizer.lbfgs.line_search_j" -LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_K = "optimizer.lbfgs.line_search_k" -LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_END = "optimizer.lbfgs.line_search_end" -LLM_KV_OPTIMIZER_LBFGS_NO_IMPROVEMENT_COUNT = "optimizer.lbfgs.no_improvement_count" - -LLM_TENSOR_OPTIMIZER_ADAM_FIRST_MOMENTS = "optimizer.adam.first_moments" -LLM_TENSOR_OPTIMIZER_ADAM_SECOND_MOMENTS = "optimizer.adam.second_moments" -LLM_TENSOR_OPTIMIZER_ADAM_PAST_LOSS_VALUES = "optimizer.adam.past_loss_values" - -LLM_TENSOR_OPTIMIZER_LBFGS_CURRENT_PARAMETERS = "optimizer.lbfgs.current_parameters" -LLM_TENSOR_OPTIMIZER_LBFGS_PREVIOUS_PARAMETERS = "optimizer.lbfgs.previous_parameters" -LLM_TENSOR_OPTIMIZER_LBFGS_CURRENT_GRADIENTS = "optimizer.lbfgs.current_gradients" -LLM_TENSOR_OPTIMIZER_LBFGS_PREVIOUS_GRADIENTS = "optimizer.lbfgs.previous_gradients" -LLM_TENSOR_OPTIMIZER_LBFGS_SEARCH_DIRECTION = "optimizer.lbfgs.search_direction" -LLM_TENSOR_OPTIMIZER_LBFGS_PAST_LOSS_VALUES = "optimizer.lbfgs.past_loss_values" -LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_ALPHA = "optimizer.lbfgs.memory_alpha" -LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_YS = "optimizer.lbfgs.memory_ys" -LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_S = "optimizer.lbfgs.memory_s" -LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_Y = "optimizer.lbfgs.memory_y" - -LLM_KV_TRAINING_TYPE_TRAIN_MODEL = "train_model" -LLM_KV_TRAINING_TYPE_FINETUNE_LORA = "finetune_lora" -LLM_KV_TRAINING_TYPE = "training.type" -LLM_KV_TRAINING_FILE_VERSION = "training.file_version" -LLM_KV_TRAINING_ITERATION_COUNT = "training.iteration_count" -LLM_KV_TRAINING_SAMPLE_COUNT = "training.sample_count" -LLM_KV_TRAINING_TOKEN_COUNT = "training.token_count" - -LLM_KV_TRAINING_LORA_RANK_TOKEN_EMBD = "training.lora.rank.token_embd" -LLM_KV_TRAINING_LORA_RANK_OUTPUT_NORM = "training.lora.rank.output_norm" -LLM_KV_TRAINING_LORA_RANK_OUTPUT = "training.lora.rank.output" -LLM_KV_TRAINING_LORA_RANK_ATTN_NORM = "training.lora.rank.attn_norm" -LLM_KV_TRAINING_LORA_RANK_ATTN_Q = "training.lora.rank.attn_q" -LLM_KV_TRAINING_LORA_RANK_ATTN_K = "training.lora.rank.attn_k" -LLM_KV_TRAINING_LORA_RANK_ATTN_V = "training.lora.rank.attn_v" -LLM_KV_TRAINING_LORA_RANK_ATTN_OUT = "training.lora.rank.attn_output" -LLM_KV_TRAINING_LORA_RANK_FFN_NORM = "training.lora.rank.ffn_norm" -LLM_KV_TRAINING_LORA_RANK_FFN_GATE = "training.lora.rank.ffn_gate" -LLM_KV_TRAINING_LORA_RANK_FFN_DOWN = "training.lora.rank.ffn_down" -LLM_KV_TRAINING_LORA_RANK_FFN_UP = "training.lora.rank.ffn_up" - -class Tensor: - def __init__(self, dtype='f', ne=None): - if ne is None: - ne = [] - self.dtype = dtype - self.ne = ne - self.nbytes = 0 - if self.dtype == 'f': - if len(self.ne) == 0: - self.nbytes = 0 - else: - self.nbytes = int(np.product(self.ne)) * 4 - else: - raise ValueError(f"Unhandled data type '{self.dtype}'") - - def load(self, data, offset): - nd = struct.unpack(' 0 else []) - - self.lbfgs_x = Tensor('f', [self.nx]) - self.lbfgs_xp = Tensor('f', [self.nx]) - self.lbfgs_g = Tensor('f', [self.nx]) - self.lbfgs_gp = Tensor('f', [self.nx]) - self.lbfgs_d = Tensor('f', [self.nx]) - self.lbfgs_pf = Tensor('f', [self.past] if self.past > 0 else []) - self.lbfgs_lmal = Tensor('f', [self.lbfgs_m]) - self.lbfgs_lmys = Tensor('f', [self.lbfgs_m]) - self.lbfgs_lms = Tensor('f', [self.nx, self.lbfgs_m]) - self.lbfgs_lmy = Tensor('f', [self.nx, self.lbfgs_m]) - - # forgot to save type in version 1: - # guess self.type from number of remaining bytes - size_type_0 = 12 + sum([t.max_storage_size() for t in - [self.adam_m, self.adam_v] - +([self.adam_pf] if (self.past > 0) else [])]) - size_type_1 = 24 + sum([t.max_storage_size() for t in - [self.lbfgs_x, self.lbfgs_xp, self.lbfgs_g, - self.lbfgs_gp, self.lbfgs_d, self.lbfgs_pf, - self.lbfgs_lmal, self.lbfgs_lmys, - self.lbfgs_lms, self.lbfgs_lmy] - +([self.lbfgs_pf] if (self.past > 0) else [])]) - # due to alignment padding the size might not by exact - # but the difference in size for both types is significant, - # so we can just use whichever is closest - remaining = len(data) - offset - if abs(remaining - size_type_0) < abs(remaining - size_type_1): - self.type = 0 - else: - self.type = 1 - - if self.type == 0: - offset = self.adam_m.load(data, offset) - offset = self.adam_v.load(data, offset) - offset = self.adam_pf.load(data,offset) - - self.adam_fx_best = struct.unpack(' 0: - self.adam_pf.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_ADAM_PAST_LOSS_VALUES) - - elif self.type == 1: - gguf_writer.add_string(LLM_KV_OPTIMIZER_TYPE, LLM_KV_OPTIMIZER_TYPE_LBFGS) - gguf_writer.add_uint32(LLM_KV_OPTIMIZER_LBFGS_APPROX_HESSIAN_COUNT, self.lbfgs_m) - gguf_writer.add_float32(LLM_KV_OPTIMIZER_LBFGS_BEST_LOSS, self.lbfgs_fx_best) - gguf_writer.add_float32(LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_STEP, self.lbfgs_step) - gguf_writer.add_int32(LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_J, self.lbfgs_j) - gguf_writer.add_int32(LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_K, self.lbfgs_k) - gguf_writer.add_int32(LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_END, self.lbfgs_end) - gguf_writer.add_uint32(LLM_KV_OPTIMIZER_LBFGS_NO_IMPROVEMENT_COUNT, self.lbfgs_n_no_improvement) - - self.lbfgs_x.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_LBFGS_CURRENT_PARAMETERS) - self.lbfgs_xp.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_LBFGS_PREVIOUS_PARAMETERS) - self.lbfgs_g.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_LBFGS_CURRENT_GRADIENTS) - self.lbfgs_gp.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_LBFGS_PREVIOUS_GRADIENTS) - self.lbfgs_d.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_LBFGS_SEARCH_DIRECTION) - if self.past > 0: - self.lbfgs_pf.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_LBFGS_PAST_LOSS_VALUES) - self.lbfgs_lmal.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_ALPHA) - self.lbfgs_lmys.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_YS) - self.lbfgs_lms.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_S) - self.lbfgs_lmy.save_gguf(gguf_writer, name=LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_Y) - else: - raise ValueError('Unknown optimizer type') - -class LoraParams: - def __init__(self): - pass - - def load(self, data, offset): - self.n_rank_attention_norm = struct.unpack(' -#include -#include -#include -#include - -#if defined(_MSC_VER) -#pragma warning(disable: 4244 4267) // possible loss of data -#endif - -struct my_llama_hparams { - uint32_t n_vocab = 32000; - uint32_t n_ctx = 512; - uint32_t n_embd = 4096; - uint32_t n_ff = 11008; - uint32_t n_head = 32; - uint32_t n_head_kv = 32; - uint32_t n_layer = 32; - - // float f_norm_eps = 1e-5f; // falcon - float f_norm_rms_eps = 1e-5f; // llama - - float rope_freq_base = 10000.0f; - float rope_freq_scale = 1.0f; - - uint32_t n_gqa() const { - return n_head/n_head_kv; - } - - uint32_t n_embd_head() const { - return n_embd/n_head; - } - - uint32_t n_embd_gqa() const { - return n_embd/n_gqa(); - } - - bool operator!=(const my_llama_hparams& other) const { - return memcmp(this, &other, sizeof(other)); - } -}; - -struct my_llama_layer { - // normalization - struct ggml_tensor * attention_norm; - - // attention - struct ggml_tensor * wq; - struct ggml_tensor * wk; - struct ggml_tensor * wv; - struct ggml_tensor * wo; - - // normalization - struct ggml_tensor * ffn_norm; - - // ff - struct ggml_tensor * ffn_gate; // w1 - struct ggml_tensor * ffn_down; // w2 - struct ggml_tensor * ffn_up; // w3 -}; - -struct my_llama_model { - struct my_llama_hparams hparams; - - struct ggml_tensor * tok_embeddings; - - struct ggml_tensor * norm; - struct ggml_tensor * output; - - std::vector layers; -}; - -struct my_llama_lora_hparams { - uint32_t lora_r = 1; - uint32_t lora_alpha = 1; - uint32_t n_rank_attention_norm = 1; - uint32_t n_rank_wq = 4; - uint32_t n_rank_wk = 4; - uint32_t n_rank_wv = 4; - uint32_t n_rank_wo = 4; - uint32_t n_rank_ffn_norm = 1; - uint32_t n_rank_ffn_gate = 4; - uint32_t n_rank_ffn_down = 4; - uint32_t n_rank_ffn_up = 4; - uint32_t n_rank_tok_embeddings = 4; - uint32_t n_rank_norm = 1; - uint32_t n_rank_output = 4; - - bool operator!=(const my_llama_lora_hparams& other) const { - return memcmp(this, &other, sizeof(other)); - } -}; - -struct my_llama_lora_layer { - // normalization - struct ggml_tensor * attention_norm_a; - struct ggml_tensor * attention_norm_b; - - // attention - struct ggml_tensor * wq_a; - struct ggml_tensor * wq_b; - struct ggml_tensor * wk_a; - struct ggml_tensor * wk_b; - struct ggml_tensor * wv_a; - struct ggml_tensor * wv_b; - struct ggml_tensor * wo_a; - struct ggml_tensor * wo_b; - - // normalization - struct ggml_tensor * ffn_norm_a; - struct ggml_tensor * ffn_norm_b; - - // ff - struct ggml_tensor * ffn_gate_a; - struct ggml_tensor * ffn_gate_b; - struct ggml_tensor * ffn_down_a; - struct ggml_tensor * ffn_down_b; - struct ggml_tensor * ffn_up_a; - struct ggml_tensor * ffn_up_b; -}; - -struct my_llama_lora { - struct ggml_context * ctx = NULL; - ggml_backend_buffer_t data; - - my_llama_lora_hparams hparams; - - struct ggml_tensor * tok_embeddings_a; - struct ggml_tensor * tok_embeddings_b; - - struct ggml_tensor * norm_a; - struct ggml_tensor * norm_b; - struct ggml_tensor * output_a; - struct ggml_tensor * output_b; - - std::vector layers; -}; - -// gguf constants -static const char * LLM_KV_TRAINING_TYPE_FINETUNE_LORA = "finetune_lora"; -static const char * LLM_KV_TRAINING_TYPE = "training.type"; - -static const char * LLM_KV_TRAINING_LORA_RANK_TOKEN_EMBD = "training.lora.rank.token_embd"; -static const char * LLM_KV_TRAINING_LORA_RANK_OUTPUT_NORM = "training.lora.rank.output_norm"; -static const char * LLM_KV_TRAINING_LORA_RANK_OUTPUT = "training.lora.rank.output"; -static const char * LLM_KV_TRAINING_LORA_RANK_ATTN_NORM = "training.lora.rank.attn_norm"; -static const char * LLM_KV_TRAINING_LORA_RANK_ATTN_Q = "training.lora.rank.attn_q"; -static const char * LLM_KV_TRAINING_LORA_RANK_ATTN_K = "training.lora.rank.attn_k"; -static const char * LLM_KV_TRAINING_LORA_RANK_ATTN_V = "training.lora.rank.attn_v"; -static const char * LLM_KV_TRAINING_LORA_RANK_ATTN_OUT = "training.lora.rank.attn_output"; -static const char * LLM_KV_TRAINING_LORA_RANK_FFN_NORM = "training.lora.rank.ffn_norm"; -static const char * LLM_KV_TRAINING_LORA_RANK_FFN_GATE = "training.lora.rank.ffn_gate"; -static const char * LLM_KV_TRAINING_LORA_RANK_FFN_DOWN = "training.lora.rank.ffn_down"; -static const char * LLM_KV_TRAINING_LORA_RANK_FFN_UP = "training.lora.rank.ffn_up"; - -// gguf constants (sync with gguf.py) - -static const char * LLM_KV_GENERAL_ARCHITECTURE = "general.architecture"; -static const char * LLM_KV_GENERAL_FILE_TYPE = "general.file_type"; - -static const char * LLM_KV_CONTEXT_LENGTH = "%s.context_length"; -static const char * LLM_KV_EMBEDDING_LENGTH = "%s.embedding_length"; -static const char * LLM_KV_BLOCK_COUNT = "%s.block_count"; -static const char * LLM_KV_FEED_FORWARD_LENGTH = "%s.feed_forward_length"; -static const char * LLM_KV_ATTENTION_HEAD_COUNT = "%s.attention.head_count"; -static const char * LLM_KV_ATTENTION_HEAD_COUNT_KV = "%s.attention.head_count_kv"; -static const char * LLM_KV_ATTENTION_LAYERNORM_RMS_EPS = "%s.attention.layer_norm_rms_epsilon"; -static const char * LLM_KV_ROPE_DIMENSION_COUNT = "%s.rope.dimension_count"; -static const char * LLM_KV_ROPE_FREQ_BASE = "%s.rope.freq_base"; // TODO load in llama.cpp -static const char * LLM_KV_ROPE_SCALE_LINEAR = "%s.rope.scale_linear"; - -static const char * LLM_TENSOR_TOKEN_EMBD = "token_embd"; -static const char * LLM_TENSOR_OUTPUT_NORM = "output_norm"; -static const char * LLM_TENSOR_OUTPUT = "output"; -static const char * LLM_TENSOR_ATTN_NORM = "blk.%d.attn_norm"; -static const char * LLM_TENSOR_ATTN_Q = "blk.%d.attn_q"; -static const char * LLM_TENSOR_ATTN_K = "blk.%d.attn_k"; -static const char * LLM_TENSOR_ATTN_V = "blk.%d.attn_v"; -static const char * LLM_TENSOR_ATTN_OUT = "blk.%d.attn_output"; -static const char * LLM_TENSOR_FFN_NORM = "blk.%d.ffn_norm"; -static const char * LLM_TENSOR_FFN_GATE = "blk.%d.ffn_gate"; -static const char * LLM_TENSOR_FFN_DOWN = "blk.%d.ffn_down"; -static const char * LLM_TENSOR_FFN_UP = "blk.%d.ffn_up"; - -static void print_params(struct my_llama_hparams * params) { - printf("%s: n_vocab : %u\n", __func__, params->n_vocab); - printf("%s: n_ctx : %u\n", __func__, params->n_ctx); - printf("%s: n_embd : %u\n", __func__, params->n_embd); - printf("%s: n_ff : %u\n", __func__, params->n_ff); - printf("%s: n_head : %u\n", __func__, params->n_head); - printf("%s: n_head_kv : %u\n", __func__, params->n_head_kv); - printf("%s: n_layer : %u\n", __func__, params->n_layer); - printf("%s: norm_rms_eps : %f\n", __func__, params->f_norm_rms_eps); - printf("%s: rope_freq_base : %f\n", __func__, params->rope_freq_base); - printf("%s: rope_freq_scale : %f\n", __func__, params->rope_freq_scale); -} - -static void print_lora_params(struct my_llama_lora_hparams * params) { - printf("%s: n_rank_attention_norm : %u\n", __func__, params->n_rank_attention_norm); - printf("%s: n_rank_wq : %u\n", __func__, params->n_rank_wq); - printf("%s: n_rank_wk : %u\n", __func__, params->n_rank_wk); - printf("%s: n_rank_wv : %u\n", __func__, params->n_rank_wv); - printf("%s: n_rank_wo : %u\n", __func__, params->n_rank_wo); - printf("%s: n_rank_ffn_norm : %u\n", __func__, params->n_rank_ffn_norm); - printf("%s: n_rank_ffn_gate : %u\n", __func__, params->n_rank_ffn_gate); - printf("%s: n_rank_ffn_down : %u\n", __func__, params->n_rank_ffn_down); - printf("%s: n_rank_ffn_up : %u\n", __func__, params->n_rank_ffn_up); - printf("%s: n_rank_tok_embeddings : %u\n", __func__, params->n_rank_tok_embeddings); - printf("%s: n_rank_norm : %u\n", __func__, params->n_rank_norm); - printf("%s: n_rank_output : %u\n", __func__, params->n_rank_output); -} - -#define GGUF_GET_KEY(ctx, dst, func, type, req, key) \ -{ \ - const std::string skey(key); \ - const int kid = gguf_find_key(ctx, skey.c_str()); \ - if (kid >= 0) { \ - enum gguf_type ktype = gguf_get_kv_type(ctx, kid); \ - if (ktype != (type)) { \ - die_fmt("key %s has wrong type: %s", skey.c_str(), gguf_type_name(ktype)); \ - } \ - (dst) = func(ctx, kid); \ - } else if (req) { \ - die_fmt("key not found in model: %s", skey.c_str()); \ - } \ -} - -static void load_model_hparams_gguf(struct gguf_context * ctx, struct my_llama_hparams * hparams, const char * expected_arch) { - std::string arch; - - GGUF_GET_KEY(ctx, arch, gguf_get_val_str, GGUF_TYPE_STRING, true, LLM_KV_GENERAL_ARCHITECTURE); - if (expected_arch != NULL) { - if (arch != expected_arch) { - printf("%s: arch=%s expected_arch=%s\n", __func__, arch.c_str(), expected_arch); - } - GGML_ASSERT(arch == expected_arch); - } - - std::vector keybuf; - keybuf.resize(512); - auto kv = [&arch, &keybuf](const char * key) -> const char * { - snprintf(keybuf.data(), keybuf.size(), key, arch.c_str()); - return keybuf.data(); - }; - - GGUF_GET_KEY(ctx, hparams->n_embd, gguf_get_val_u32, GGUF_TYPE_UINT32, true, kv(LLM_KV_EMBEDDING_LENGTH)); - GGUF_GET_KEY(ctx, hparams->n_ctx, gguf_get_val_u32, GGUF_TYPE_UINT32, false, kv(LLM_KV_CONTEXT_LENGTH)); - GGUF_GET_KEY(ctx, hparams->n_ff, gguf_get_val_u32, GGUF_TYPE_UINT32, true, kv(LLM_KV_FEED_FORWARD_LENGTH)); - GGUF_GET_KEY(ctx, hparams->n_head, gguf_get_val_u32, GGUF_TYPE_UINT32, true, kv(LLM_KV_ATTENTION_HEAD_COUNT)); - GGUF_GET_KEY(ctx, hparams->n_layer, gguf_get_val_u32, GGUF_TYPE_UINT32, true, kv(LLM_KV_BLOCK_COUNT)); - - // n_head_kv is optional, default to n_head - hparams->n_head_kv = hparams->n_head; - GGUF_GET_KEY(ctx, hparams->n_head_kv, gguf_get_val_u32, GGUF_TYPE_UINT32, false, kv(LLM_KV_ATTENTION_HEAD_COUNT_KV)); - - float rope_freq_scale = 1.0f; - GGUF_GET_KEY(ctx, hparams->f_norm_rms_eps, gguf_get_val_f32, GGUF_TYPE_FLOAT32, false, kv(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS)); - GGUF_GET_KEY(ctx, hparams->rope_freq_base, gguf_get_val_f32, GGUF_TYPE_FLOAT32, false, kv(LLM_KV_ROPE_FREQ_BASE)); - GGUF_GET_KEY(ctx, rope_freq_scale, gguf_get_val_f32, GGUF_TYPE_FLOAT32, false, kv(LLM_KV_ROPE_SCALE_LINEAR)); - if (rope_freq_scale != 1.0f) { - hparams->rope_freq_scale = 1.0f / rope_freq_scale; - } -} - -static void init_model(struct llama_model * input, struct my_llama_model * model, const char * fn_model, uint32_t n_ctx) { - auto & hparams = model->hparams; - - std::vector tn_buf; - tn_buf.resize(GGML_MAX_NAME); - auto tn = [&tn_buf](const char * key) -> const char * { - snprintf(tn_buf.data(), tn_buf.size(), "%s.weight", key); - return tn_buf.data(); - }; - auto tni = [&tn_buf](const char * key, int bid) -> const char * { - snprintf(tn_buf.data(), tn_buf.size(), key, bid); - std::string s = tn_buf.data(); - snprintf(tn_buf.data(), tn_buf.size(), "%s.weight", s.c_str()); - return tn_buf.data(); - }; - - - // get parameters directly from gguf file - { - struct gguf_init_params params = { - /*.no_alloc = */ false, - /*.ctx = */ NULL, - }; - struct gguf_context * mctx = gguf_init_from_file(fn_model, params); - - load_model_hparams_gguf(mctx, &hparams, "llama"); - - gguf_free(mctx); - } - hparams.n_vocab = llama_n_vocab(input); - hparams.n_ctx = n_ctx; - - // get tensors from llama_model (possibly mmapped) - model->tok_embeddings = llama_get_model_tensor(input, tn(LLM_TENSOR_TOKEN_EMBD)); - model->norm = llama_get_model_tensor(input, tn(LLM_TENSOR_OUTPUT_NORM)); - model->output = llama_get_model_tensor(input, tn(LLM_TENSOR_OUTPUT)); - - assert_shape_2d(model->tok_embeddings, hparams.n_embd, hparams.n_vocab); - assert_shape_1d(model->norm, hparams.n_embd); - assert_shape_2d(model->output, hparams.n_embd, hparams.n_vocab); - - model->layers.resize(hparams.n_layer); - for (uint32_t i = 0; i < hparams.n_layer; ++i) { - auto & layer = model->layers[i]; - - layer.attention_norm = llama_get_model_tensor(input, tni(LLM_TENSOR_ATTN_NORM, i)); - layer.wq = llama_get_model_tensor(input, tni(LLM_TENSOR_ATTN_Q, i)); - layer.wk = llama_get_model_tensor(input, tni(LLM_TENSOR_ATTN_K, i)); - layer.wv = llama_get_model_tensor(input, tni(LLM_TENSOR_ATTN_V, i)); - layer.wo = llama_get_model_tensor(input, tni(LLM_TENSOR_ATTN_OUT, i)); - layer.ffn_norm = llama_get_model_tensor(input, tni(LLM_TENSOR_FFN_NORM, i)); - layer.ffn_gate = llama_get_model_tensor(input, tni(LLM_TENSOR_FFN_GATE, i)); - layer.ffn_down = llama_get_model_tensor(input, tni(LLM_TENSOR_FFN_DOWN, i)); - layer.ffn_up = llama_get_model_tensor(input, tni(LLM_TENSOR_FFN_UP, i)); - - assert_shape_1d(layer.attention_norm, hparams.n_embd); - assert_shape_2d(layer.wq, hparams.n_embd, hparams.n_embd); - assert_shape_2d(layer.wk, hparams.n_embd, hparams.n_embd_gqa()); - assert_shape_2d(layer.wv, hparams.n_embd, hparams.n_embd_gqa()); - assert_shape_2d(layer.wo, hparams.n_embd, hparams.n_embd); - assert_shape_1d(layer.ffn_norm, hparams.n_embd); - assert_shape_2d(layer.ffn_gate, hparams.n_embd, hparams.n_ff); - assert_shape_2d(layer.ffn_down, hparams.n_ff, hparams.n_embd); - assert_shape_2d(layer.ffn_up, hparams.n_embd, hparams.n_ff); - } -} - -static void set_param_lora(struct my_llama_lora * lora) { - const uint32_t n_layer = lora->layers.size(); - - struct ggml_context* ctx = lora->ctx; - - ggml_set_param(ctx, lora->tok_embeddings_a); - ggml_set_param(ctx, lora->tok_embeddings_b); - ggml_set_param(ctx, lora->norm_a); - ggml_set_param(ctx, lora->norm_b); - ggml_set_param(ctx, lora->output_a); - ggml_set_param(ctx, lora->output_b); - - for (uint32_t i = 0; i < n_layer; ++i) { - auto & layer = lora->layers[i]; - - ggml_set_param(ctx, layer.attention_norm_a); - ggml_set_param(ctx, layer.attention_norm_b); - ggml_set_param(ctx, layer.wq_a); - ggml_set_param(ctx, layer.wq_b); - ggml_set_param(ctx, layer.wk_a); - ggml_set_param(ctx, layer.wk_b); - ggml_set_param(ctx, layer.wv_a); - ggml_set_param(ctx, layer.wv_b); - ggml_set_param(ctx, layer.wo_a); - ggml_set_param(ctx, layer.wo_b); - ggml_set_param(ctx, layer.ffn_norm_a); - ggml_set_param(ctx, layer.ffn_norm_b); - ggml_set_param(ctx, layer.ffn_gate_a); - ggml_set_param(ctx, layer.ffn_gate_b); - ggml_set_param(ctx, layer.ffn_down_a); - ggml_set_param(ctx, layer.ffn_down_b); - ggml_set_param(ctx, layer.ffn_up_a); - ggml_set_param(ctx, layer.ffn_up_b); - } -} - -static void init_lora(const struct my_llama_model * model, struct my_llama_lora * lora) { - const auto & lparams = lora->hparams; - - const uint32_t n_embd = model->hparams.n_embd; - const uint32_t n_embd_gqa = model->hparams.n_embd_gqa(); - const uint32_t n_layer = model->hparams.n_layer; - const uint32_t n_vocab = model->hparams.n_vocab; - const uint32_t n_ff = model->hparams.n_ff; - - std::vector tn_buf; - tn_buf.resize(GGML_MAX_NAME); - auto tn = [&tn_buf](const char * key, const char * suffix) -> const char * { - snprintf(tn_buf.data(), tn_buf.size(), "%s%s", key, suffix); - return tn_buf.data(); - }; - auto tni = [&tn_buf](const char * key, const char * suffix, int bid) -> const char * { - snprintf(tn_buf.data(), tn_buf.size(), key, bid); - std::string s = tn_buf.data(); - snprintf(tn_buf.data(), tn_buf.size(), "%s%s", s.c_str(), suffix); - return tn_buf.data(); - }; - - // context for lora tensors without their data - struct ggml_init_params ctx_lora_params; - ctx_lora_params.mem_size = ggml_tensor_overhead()*2*(6 + n_layer*18); - ctx_lora_params.mem_buffer = NULL; - ctx_lora_params.no_alloc = true; - - struct ggml_context * ctx = ggml_init(ctx_lora_params); - lora->ctx = ctx; - - lora->tok_embeddings_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_tok_embeddings, n_embd); - lora->tok_embeddings_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_tok_embeddings, n_vocab); - lora->norm_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_norm, n_embd); - lora->norm_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_norm, 1); - lora->output_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_output, n_embd); - lora->output_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_output, n_vocab); - - ggml_set_name(lora->tok_embeddings_a, tn(LLM_TENSOR_TOKEN_EMBD, ".weight.lora_a")); - ggml_set_name(lora->tok_embeddings_b, tn(LLM_TENSOR_TOKEN_EMBD, ".weight.lora_b")); - ggml_set_name(lora->norm_a, tn(LLM_TENSOR_OUTPUT_NORM, ".weight.lora_a")); - ggml_set_name(lora->norm_b, tn(LLM_TENSOR_OUTPUT_NORM, ".weight.lora_b")); - ggml_set_name(lora->output_a, tn(LLM_TENSOR_OUTPUT, ".weight.lora_a")); - ggml_set_name(lora->output_b, tn(LLM_TENSOR_OUTPUT, ".weight.lora_b")); - - lora->layers.resize(n_layer); - for (uint32_t i = 0; i < n_layer; ++i) { - auto & layer = lora->layers[i]; - - layer.attention_norm_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_attention_norm, n_embd); - layer.attention_norm_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_attention_norm, 1); - - layer.wq_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_wq, n_embd); - layer.wq_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_wq, n_embd); - layer.wk_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_wk, n_embd); - layer.wk_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_wk, n_embd_gqa); - layer.wv_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_wv, n_embd); - layer.wv_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_wv, n_embd_gqa); - layer.wo_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_wo, n_embd); - layer.wo_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_wo, n_embd); - - layer.ffn_norm_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_ffn_norm, n_embd); - layer.ffn_norm_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_ffn_norm, 1); - - layer.ffn_gate_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_ffn_gate, n_embd); - layer.ffn_gate_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_ffn_gate, n_ff); - layer.ffn_down_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_ffn_down, n_ff); - layer.ffn_down_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_ffn_down, n_embd); - layer.ffn_up_a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_ffn_up, n_embd); - layer.ffn_up_b = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, lparams.n_rank_ffn_up, n_ff); - - ggml_set_name(layer.attention_norm_a, tni(LLM_TENSOR_ATTN_NORM, ".weight.lora_a", i)); - ggml_set_name(layer.attention_norm_b, tni(LLM_TENSOR_ATTN_NORM, ".weight.lora_b", i)); - ggml_set_name(layer.wq_a, tni(LLM_TENSOR_ATTN_Q, ".weight.lora_a", i)); - ggml_set_name(layer.wq_b, tni(LLM_TENSOR_ATTN_Q, ".weight.lora_b", i)); - ggml_set_name(layer.wk_a, tni(LLM_TENSOR_ATTN_K, ".weight.lora_a", i)); - ggml_set_name(layer.wk_b, tni(LLM_TENSOR_ATTN_K, ".weight.lora_b", i)); - ggml_set_name(layer.wv_a, tni(LLM_TENSOR_ATTN_V, ".weight.lora_a", i)); - ggml_set_name(layer.wv_b, tni(LLM_TENSOR_ATTN_V, ".weight.lora_b", i)); - ggml_set_name(layer.wo_a, tni(LLM_TENSOR_ATTN_OUT, ".weight.lora_a", i)); - ggml_set_name(layer.wo_b, tni(LLM_TENSOR_ATTN_OUT, ".weight.lora_b", i)); - ggml_set_name(layer.ffn_norm_a, tni(LLM_TENSOR_FFN_NORM, ".weight.lora_a", i)); - ggml_set_name(layer.ffn_norm_b, tni(LLM_TENSOR_FFN_NORM, ".weight.lora_b", i)); - ggml_set_name(layer.ffn_gate_a, tni(LLM_TENSOR_FFN_GATE, ".weight.lora_a", i)); - ggml_set_name(layer.ffn_gate_b, tni(LLM_TENSOR_FFN_GATE, ".weight.lora_b", i)); - ggml_set_name(layer.ffn_down_a, tni(LLM_TENSOR_FFN_DOWN, ".weight.lora_a", i)); - ggml_set_name(layer.ffn_down_b, tni(LLM_TENSOR_FFN_DOWN, ".weight.lora_b", i)); - ggml_set_name(layer.ffn_up_a, tni(LLM_TENSOR_FFN_UP, ".weight.lora_a", i)); - ggml_set_name(layer.ffn_up_b, tni(LLM_TENSOR_FFN_UP, ".weight.lora_b", i)); - } - - set_param_lora(lora); - - // allocate data for lora tensors - lora->data = ggml_backend_alloc_ctx_tensors_from_buft(ctx, ggml_backend_cpu_buffer_type()); -} - -static void randomize_lora(struct my_llama_lora * lora, int seed, float mean, float std, float min, float max) { - const uint32_t n_layer = lora->layers.size(); - - struct random_normal_distribution * rnd = init_random_normal_distribution(seed, mean, std, min, max); - - randomize_tensor_normal(lora->tok_embeddings_a, rnd); - ggml_set_zero(lora->tok_embeddings_b); - randomize_tensor_normal(lora->norm_a, rnd); - ggml_set_zero(lora->norm_b); - randomize_tensor_normal(lora->output_a, rnd); - ggml_set_zero(lora->output_b); - - for (uint32_t i = 0; i < n_layer; ++i) { - auto & layer = lora->layers[i]; - randomize_tensor_normal(layer.attention_norm_a, rnd); - ggml_set_zero(layer.attention_norm_b); - - randomize_tensor_normal(layer.wq_a, rnd); - ggml_set_zero(layer.wq_b); - randomize_tensor_normal(layer.wk_a, rnd); - ggml_set_zero(layer.wk_b); - randomize_tensor_normal(layer.wv_a, rnd); - ggml_set_zero(layer.wv_b); - randomize_tensor_normal(layer.wo_a, rnd); - ggml_set_zero(layer.wo_b); - - randomize_tensor_normal(layer.ffn_norm_a, rnd); - ggml_set_zero(layer.ffn_norm_b); - - randomize_tensor_normal(layer.ffn_gate_a, rnd); - ggml_set_zero(layer.ffn_gate_b); - randomize_tensor_normal(layer.ffn_down_a, rnd); - ggml_set_zero(layer.ffn_down_b); - randomize_tensor_normal(layer.ffn_up_a, rnd); - ggml_set_zero(layer.ffn_up_b); - } - - free_random_normal_distribution(rnd); -} - -static struct ggml_tensor * llama_build_lora_finetune_graphs( - struct my_llama_model * model, - struct my_llama_lora * lora, - ggml_gallocr_t alloc, - struct ggml_context * ctx, - struct ggml_cgraph * gf, - struct ggml_cgraph * gb, - struct ggml_cgraph * gb_tmp, - struct ggml_tensor * * logits, - struct ggml_tensor * tokens_input, - struct ggml_tensor * targets, - const int n_tokens, - const int n_batch, - const bool enable_flash_attn, - const bool enable_checkpointing, - const bool measure_only) { - - ggml_set_scratch(ctx, { 0, 0, nullptr, }); - const int n_past = 0; - const int N = n_tokens; - const auto & hparams = model->hparams; - const int n_ctx = hparams.n_ctx; - const int n_vocab = hparams.n_vocab; - const int n_embd = hparams.n_embd; - const int n_layer = hparams.n_layer; - const int n_head = hparams.n_head; - const int n_head_kv = hparams.n_head_kv; - const int n_ff = hparams.n_ff; - const int n_rot = hparams.n_embd_head(); - const int n_embd_head = hparams.n_embd_head(); - const int n_embd_gqa = hparams.n_embd_gqa(); - - const float rms_norm_eps = hparams.f_norm_rms_eps; - const float rope_freq_base = hparams.rope_freq_base; - const float rope_freq_scale = hparams.rope_freq_scale; - - GGML_ASSERT((size_t) n_layer == lora->layers.size()); - - auto set_name = [](struct ggml_tensor * t, const char * n) { - ggml_set_name(t, n); - if (t->grad) { - ggml_format_name(t->grad, "%s->grad", n); - } - }; - - // KQ_pos - contains the positions - struct ggml_tensor * KQ_pos = ggml_new_tensor_1d(ctx, GGML_TYPE_I32, N); - ggml_set_input(KQ_pos); - - // rope has so much parameters that we make a custom function for it - auto rope = [ctx, KQ_pos, n_rot, n_ctx, rope_freq_base, rope_freq_scale] - (struct ggml_tensor * t) -> struct ggml_tensor * { - // not capturing these, to silcence warnings - const int rope_mode = 0; - - return ggml_rope_ext(ctx, - t, KQ_pos, nullptr, n_rot, rope_mode, n_ctx, - rope_freq_base, rope_freq_scale, 0.0f, 1.0f, 0.0f, 0.0f - ); - }; - - set_name(tokens_input, "tokens_input"); - set_name(targets, "targets"); - - GGML_ASSERT(tokens_input->type == GGML_TYPE_I32); - - auto add_to_f32 = [] (struct ggml_context * ctx, struct ggml_tensor * a, struct ggml_tensor * b) { - if (ggml_is_quantized(a->type) || a->type == GGML_TYPE_F16 || a->type == GGML_TYPE_BF16) { - return ggml_add_cast(ctx, a, b, GGML_TYPE_F32); - } else if (a->type == GGML_TYPE_F32) { - return ggml_add(ctx, a, b); - } else { - die_fmt("%s: Finetuning on tensors with type '%s' is not yet supported.\n", - __func__, ggml_type_name(a->type)); - } - }; - - struct ggml_tensor * tok_embeddings = add_to_f32(ctx, model->tok_embeddings, ggml_mul_mat(ctx, lora->tok_embeddings_a, lora->tok_embeddings_b)); - struct ggml_tensor * norm = add_to_f32(ctx, model->norm, ggml_mul_mat(ctx, lora->norm_a, lora->norm_b)); - struct ggml_tensor * output = add_to_f32(ctx, model->output, ggml_mul_mat(ctx, lora->output_a, lora->output_b)); - - struct ggml_tensor * t00 = ggml_reshape_1d(ctx, tokens_input, N*n_batch); set_name(t00, "t00"); assert_shape_1d(t00, N*n_batch); - struct ggml_tensor * t01 = ggml_get_rows(ctx, tok_embeddings, t00); set_name(t01, "t01"); assert_shape_2d(t01, n_embd, N*n_batch); - - struct ggml_tensor * cur = t01; - - std::vector checkpoints; - if (enable_checkpointing) { - checkpoints.push_back(tokens_input); - checkpoints.push_back(targets); - checkpoints.push_back(t00); - checkpoints.push_back(t01); - } - - const float kv_scale = 1.0f/sqrtf(float(n_embd)/n_head); - - for (int il = 0; il < n_layer; ++il) { - struct my_llama_layer & layer = model->layers[il]; - struct my_llama_lora_layer & llayer = lora->layers[il]; - - struct ggml_tensor * attention_norm = add_to_f32(ctx, layer.attention_norm, ggml_mul_mat(ctx, llayer.attention_norm_a, llayer.attention_norm_b)); - struct ggml_tensor * ffn_norm = add_to_f32(ctx, layer.ffn_norm, ggml_mul_mat(ctx, llayer.ffn_norm_a, llayer.ffn_norm_b)); - struct ggml_tensor * wq = add_to_f32(ctx, layer.wq, ggml_mul_mat(ctx, llayer.wq_a, llayer.wq_b)); - struct ggml_tensor * wk = add_to_f32(ctx, layer.wk, ggml_mul_mat(ctx, llayer.wk_a, llayer.wk_b)); - struct ggml_tensor * wv = add_to_f32(ctx, layer.wv, ggml_mul_mat(ctx, llayer.wv_a, llayer.wv_b)); - struct ggml_tensor * wo = add_to_f32(ctx, layer.wo, ggml_mul_mat(ctx, llayer.wo_a, llayer.wo_b)); - struct ggml_tensor * ffn_gate = add_to_f32(ctx, layer.ffn_gate, ggml_mul_mat(ctx, llayer.ffn_gate_a, llayer.ffn_gate_b)); - struct ggml_tensor * ffn_down = add_to_f32(ctx, layer.ffn_down, ggml_mul_mat(ctx, llayer.ffn_down_a, llayer.ffn_down_b)); - struct ggml_tensor * ffn_up = add_to_f32(ctx, layer.ffn_up, ggml_mul_mat(ctx, llayer.ffn_up_a, llayer.ffn_up_b)); - - struct ggml_tensor * t02 = ggml_rms_norm (ctx, cur, rms_norm_eps); set_name(t02, "t02"); assert_shape_2d(t02, n_embd, N*n_batch); - struct ggml_tensor * t03 = ggml_repeat (ctx, attention_norm, t02); set_name(t03, "t03"); assert_shape_2d(t03, n_embd, N*n_batch); - struct ggml_tensor * t04 = ggml_mul (ctx, t03, t02); set_name(t04, "t04"); assert_shape_2d(t04, n_embd, N*n_batch); - struct ggml_tensor * t05 = ggml_mul_mat (ctx, wq, t04); set_name(t05, "t05"); assert_shape_2d(t05, n_embd, N*n_batch); - struct ggml_tensor * t06 = ggml_reshape_4d (ctx, t05, n_embd_head, n_head, N, n_batch); set_name(t06, "t06"); assert_shape_4d(t06, n_embd_head, n_head, N, n_batch); - struct ggml_tensor * t07 = rope (t06); set_name(t07, "t07"); assert_shape_4d(t07, n_embd_head, n_head, N, n_batch); - struct ggml_tensor * t08 = ggml_mul_mat (ctx, wk, t04); set_name(t08, "t08"); assert_shape_2d(t08, n_embd_gqa, N*n_batch); - struct ggml_tensor * t09 = ggml_reshape_4d (ctx, t08, n_embd_head, n_head_kv, N, n_batch); set_name(t09, "t09"); assert_shape_4d(t09, n_embd_head, n_head_kv, N, n_batch); - struct ggml_tensor * t10 = rope (t09); set_name(t10, "t10"); assert_shape_4d(t10, n_embd_head, n_head_kv, N, n_batch); - - struct ggml_tensor * t11; - if (ggml_is_quantized(wv->type)) { - struct ggml_tensor * t11_1 = ggml_mul_mat (ctx, wv, t04); set_name(t11_1, "t11_1"); assert_shape_2d(t11_1, n_embd_gqa, N*n_batch); - struct ggml_tensor * t11_2 = ggml_transpose(ctx, t11_1); set_name(t11_2, "t11_2"); assert_shape_2d(t11_2, N*n_batch, n_embd_gqa); - t11 = ggml_cont (ctx, t11_2); set_name(t11, "t11"); assert_shape_2d(t11, N*n_batch, n_embd_gqa); - } else { - t11 = ggml_mul_mat (ctx, t04, wv); set_name(t11, "t11"); assert_shape_2d(t11, N*n_batch, n_embd_gqa); - } - - struct ggml_tensor * t12 = ggml_reshape_4d (ctx, t11, N, n_batch, n_embd_head, n_head_kv); set_name(t12, "t12"); assert_shape_4d(t12, N, n_batch, n_embd_head, n_head_kv); - struct ggml_tensor * t13 = ggml_permute (ctx, t07, 0, 2, 1, 3); set_name(t13, "t13"); assert_shape_4d(t13, n_embd_head, N, n_head, n_batch); - struct ggml_tensor * t14 = ggml_permute (ctx, t10, 0, 2, 1, 3); set_name(t14, "t14"); assert_shape_4d(t14, n_embd_head, N, n_head_kv, n_batch); - struct ggml_tensor * t15 = ggml_permute (ctx, t12, 0, 3, 1, 2); set_name(t15, "t15"); assert_shape_4d(t15, N, n_embd_head, n_head_kv, n_batch); - struct ggml_tensor * t16; - if (enable_flash_attn) { - GGML_ASSERT(false && "TODO: ggml_flash_attn_ext() not yet supported"); - //t16 = ggml_flash_attn(ctx, t13, t14, t15, true); set_name(t16, "t16"); assert_shape_4d(t16, n_embd_head, N, n_head, n_batch); - } else { - struct ggml_tensor * t16_0 = ggml_mul_mat (ctx, t14, t13); set_name(t16_0, "t16_0"); assert_shape_4d(t16_0, N, N, n_head, n_batch); - struct ggml_tensor * t16_1 = ggml_scale_inplace (ctx, t16_0, kv_scale); set_name(t16_1, "t16_1"); assert_shape_4d(t16_1, N, N, n_head, n_batch); - struct ggml_tensor * t16_2 = ggml_diag_mask_inf_inplace(ctx, t16_1, n_past); set_name(t16_2, "t16_2"); assert_shape_4d(t16_2, N, N, n_head, n_batch); - struct ggml_tensor * t16_3 = ggml_soft_max_inplace (ctx, t16_2); set_name(t16_3, "t16_3"); assert_shape_4d(t16_3, N, N, n_head, n_batch); - t16 = ggml_mul_mat(ctx, t15, t16_3); set_name(t16, "t16"); assert_shape_4d(t16, n_embd_head, N, n_head, n_batch); - } - struct ggml_tensor * t17 = ggml_permute (ctx, t16, 0, 2, 1, 3); set_name(t17, "t17"); assert_shape_4d(t17, n_embd_head, n_head, N, n_batch); - struct ggml_tensor * t18 = ggml_cont (ctx, t17); set_name(t18, "t18"); assert_shape_4d(t18, n_embd_head, n_head, N, n_batch); - struct ggml_tensor * t19 = ggml_reshape_2d (ctx, t18, n_embd, N*n_batch); set_name(t19, "t19"); assert_shape_2d(t19, n_embd, N*n_batch); - struct ggml_tensor * t20 = ggml_mul_mat (ctx, wo, t19); set_name(t20, "t20"); assert_shape_2d(t20, n_embd, N*n_batch); - struct ggml_tensor * t21 = ggml_add (ctx, t20, cur); set_name(t21, "t21"); assert_shape_2d(t21, n_embd, N*n_batch); - struct ggml_tensor * t22 = ggml_rms_norm (ctx, t21, rms_norm_eps); set_name(t22, "t22"); assert_shape_2d(t22, n_embd, N*n_batch); - struct ggml_tensor * t23 = ggml_repeat (ctx, ffn_norm, t22); set_name(t23, "t23"); assert_shape_2d(t23, n_embd, N*n_batch); - struct ggml_tensor * t24 = ggml_mul (ctx, t23, t22); set_name(t24, "t24"); assert_shape_2d(t24, n_embd, N*n_batch); - struct ggml_tensor * t25 = ggml_mul_mat (ctx, ffn_up, t24); set_name(t25, "t25"); assert_shape_2d(t25, n_ff, N*n_batch); - struct ggml_tensor * t26 = ggml_mul_mat (ctx, ffn_gate, t24); set_name(t26, "t26"); assert_shape_2d(t26, n_ff, N*n_batch); - struct ggml_tensor * t27 = ggml_silu (ctx, t26); set_name(t27, "t27"); assert_shape_2d(t27, n_ff, N*n_batch); - struct ggml_tensor * t28 = ggml_mul (ctx, t27, t25); set_name(t28, "t28"); assert_shape_2d(t28, n_ff, N*n_batch); - struct ggml_tensor * t29 = ggml_mul_mat (ctx, ffn_down, t28); set_name(t29, "t29"); assert_shape_2d(t29, n_embd, N*n_batch); - struct ggml_tensor * t30 = ggml_add (ctx, t29, t21); set_name(t30, "t30"); assert_shape_2d(t30, n_embd, N*n_batch); - cur = t30; - if (enable_checkpointing) { - checkpoints.push_back(cur); - } - } - struct ggml_tensor * t31 = ggml_rms_norm (ctx, cur, rms_norm_eps); set_name(t31, "t31"); assert_shape_2d(t31, n_embd, N*n_batch); - struct ggml_tensor * t32 = ggml_repeat (ctx, norm, t31); set_name(t32, "t32"); assert_shape_2d(t32, n_embd, N*n_batch); - struct ggml_tensor * t33 = ggml_mul (ctx, t32, t31); set_name(t33, "t33"); assert_shape_2d(t33, n_embd, N*n_batch); - struct ggml_tensor * t34 = ggml_mul_mat (ctx, output, t33); set_name(t34, "t34"); assert_shape_2d(t34, n_vocab, N*n_batch); - struct ggml_tensor * t35 = ggml_reshape_3d (ctx, t34, n_vocab, N, n_batch); set_name(t35, "t35"); assert_shape_3d(t35, n_vocab, N, n_batch); - struct ggml_tensor * t36 = ggml_cross_entropy_loss(ctx, t35, targets); set_name(t36, "t36"); assert_shape_1d(t36, 1); - - if (enable_checkpointing) { - checkpoints.push_back(t31); - checkpoints.push_back(t32); - checkpoints.push_back(t33); - checkpoints.push_back(t34); - checkpoints.push_back(t35); - checkpoints.push_back(t36); - } - - ggml_build_forward_expand(gf, t36); - - if (enable_checkpointing) { - ggml_build_backward_gradient_checkpointing(ctx, gf, gb, gb_tmp, checkpoints.data(), (int) checkpoints.size()); - } else { - ggml_graph_cpy(gf, gb); - ggml_build_backward_expand(ctx, gf, gb, true); - } - - GGML_ASSERT(alloc != NULL); - - // make sure some tensors are not reallocated by inserting new temporary nodes depending on them - int n_leafs_before = gb->n_leafs; - int n_nodes_before = gb->n_nodes; - - // output tensors - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, t35, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, t36, 1.0f)); - // input gradient - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, t36->grad, 1.0f)); - GGML_ASSERT(t36->grad->data == NULL && t36->grad->view_src == NULL); - ggml_set_input(t36->grad); - // KQ_pos - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, KQ_pos, 1.0f)); - - // make sure base model tensors data cannot be used in viewable operations - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, model->tok_embeddings, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, model->norm, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, model->output, 1.0f)); - for (int il = 0; il < n_layer; ++il) { - struct my_llama_layer & layer = model->layers[il]; - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, layer.attention_norm, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, layer.ffn_norm, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, layer.wq, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, layer.wk, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, layer.wv, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, layer.wo, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, layer.ffn_gate, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, layer.ffn_down, 1.0f)); - ggml_build_forward_expand(gb, ggml_scale_inplace(ctx, layer.ffn_up, 1.0f)); - } - - // allocating checkpoints in one block to reduce memory fragmentation - // note: they will be freed in reverse order - for (unsigned int i = 0; i < checkpoints.size(); ++i) { - if (checkpoints[i]->data == NULL && checkpoints[i]->view_src == NULL) { - ggml_set_input(checkpoints[i]); - } - } - - if (measure_only) { - ggml_gallocr_reserve(alloc, gb); - } else { - ggml_gallocr_alloc_graph(alloc, gb); - - // set KQ_pos - { - int * data = (int *) KQ_pos->data; - for (int i = 0; i < N; ++i) { - data[i] = n_past + i; - } - } - } - - // remove the additional nodes and leafs - for (int i = n_leafs_before; i < gb->n_leafs; ++i) { - gb->leafs[i] = NULL; - } - for (int i = n_nodes_before; i < gb->n_nodes; ++i) { - gb->nodes[i] = NULL; - } - gb->n_leafs = n_leafs_before; - gb->n_nodes = n_nodes_before; - - *logits = t35; - return t36; -} - -static void load_llama_lora_gguf(struct gguf_context * fctx, struct ggml_context * f_ggml_ctx, struct my_llama_model * model, struct my_llama_lora * lora) { - // NOTE: gguf_context must be initialized with f_ggml_ctx and no_alloc=false, otherwise tensor data can not be read - - std::string arch; - - std::vector keybuf; - keybuf.resize(512); - - GGUF_GET_KEY(fctx, arch, gguf_get_val_str, GGUF_TYPE_STRING, true, LLM_KV_GENERAL_ARCHITECTURE); - GGML_ASSERT(arch == "llama"); - - uint32_t ftype_u; - GGUF_GET_KEY(fctx, ftype_u, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_GENERAL_FILE_TYPE); - GGML_ASSERT((enum llama_ftype) ftype_u == LLAMA_FTYPE_ALL_F32); - - struct my_llama_hparams hparams; - load_model_hparams_gguf(fctx, &hparams, arch.c_str()); - - // parameters that define tensor shapes must match - GGML_ASSERT(hparams.n_embd == model->hparams.n_embd); - GGML_ASSERT(hparams.n_ff == model->hparams.n_ff); - GGML_ASSERT(hparams.n_head == model->hparams.n_head); - GGML_ASSERT(hparams.n_head_kv == model->hparams.n_head_kv); - GGML_ASSERT(hparams.n_layer == model->hparams.n_layer); - - GGUF_GET_KEY(fctx, lora->hparams.n_rank_tok_embeddings, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_TOKEN_EMBD); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_norm, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_OUTPUT_NORM); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_output, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_OUTPUT); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_attention_norm, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_ATTN_NORM); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_wq, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_ATTN_Q); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_wk, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_ATTN_K); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_wv, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_ATTN_V); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_wo, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_ATTN_OUT); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_ffn_norm, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_FFN_NORM); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_ffn_gate, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_FFN_GATE); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_ffn_down, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_FFN_DOWN); - GGUF_GET_KEY(fctx, lora->hparams.n_rank_ffn_up, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_LORA_RANK_FFN_UP); - - init_lora(model, lora); - - copy_tensor_by_name(lora->tok_embeddings_a, f_ggml_ctx, ggml_get_name(lora->tok_embeddings_a)); - copy_tensor_by_name(lora->tok_embeddings_b, f_ggml_ctx, ggml_get_name(lora->tok_embeddings_b)); - copy_tensor_by_name(lora->norm_a, f_ggml_ctx, ggml_get_name(lora->norm_a)); - copy_tensor_by_name(lora->norm_b, f_ggml_ctx, ggml_get_name(lora->norm_b)); - copy_tensor_by_name(lora->output_a, f_ggml_ctx, ggml_get_name(lora->output_a)); - copy_tensor_by_name(lora->output_b, f_ggml_ctx, ggml_get_name(lora->output_b)); - - for (uint32_t i = 0; i < lora->layers.size(); ++i) { - auto & layer = lora->layers[i]; - copy_tensor_by_name(layer.attention_norm_a, f_ggml_ctx, ggml_get_name(layer.attention_norm_a)); - copy_tensor_by_name(layer.attention_norm_b, f_ggml_ctx, ggml_get_name(layer.attention_norm_b)); - copy_tensor_by_name(layer.wq_a, f_ggml_ctx, ggml_get_name(layer.wq_a)); - copy_tensor_by_name(layer.wq_b, f_ggml_ctx, ggml_get_name(layer.wq_b)); - copy_tensor_by_name(layer.wk_a, f_ggml_ctx, ggml_get_name(layer.wk_a)); - copy_tensor_by_name(layer.wk_b, f_ggml_ctx, ggml_get_name(layer.wk_b)); - copy_tensor_by_name(layer.wv_a, f_ggml_ctx, ggml_get_name(layer.wv_a)); - copy_tensor_by_name(layer.wv_b, f_ggml_ctx, ggml_get_name(layer.wv_b)); - copy_tensor_by_name(layer.wo_a, f_ggml_ctx, ggml_get_name(layer.wo_a)); - copy_tensor_by_name(layer.wo_b, f_ggml_ctx, ggml_get_name(layer.wo_b)); - copy_tensor_by_name(layer.ffn_norm_a, f_ggml_ctx, ggml_get_name(layer.ffn_norm_a)); - copy_tensor_by_name(layer.ffn_norm_b, f_ggml_ctx, ggml_get_name(layer.ffn_norm_b)); - copy_tensor_by_name(layer.ffn_gate_a, f_ggml_ctx, ggml_get_name(layer.ffn_gate_a)); - copy_tensor_by_name(layer.ffn_gate_b, f_ggml_ctx, ggml_get_name(layer.ffn_gate_b)); - copy_tensor_by_name(layer.ffn_down_a, f_ggml_ctx, ggml_get_name(layer.ffn_down_a)); - copy_tensor_by_name(layer.ffn_down_b, f_ggml_ctx, ggml_get_name(layer.ffn_down_b)); - copy_tensor_by_name(layer.ffn_up_a, f_ggml_ctx, ggml_get_name(layer.ffn_up_a)); - copy_tensor_by_name(layer.ffn_up_b, f_ggml_ctx, ggml_get_name(layer.ffn_up_b)); - } -} - -static void save_llama_lora_gguf(struct gguf_context * fctx, struct my_llama_model * model, struct my_llama_lora * lora) { - const char * arch = "llama"; - enum llama_ftype ftype = LLAMA_FTYPE_ALL_F32; - - std::vector keybuf; - keybuf.resize(512); - auto kv = [arch, &keybuf](const char * key) -> const char * { - snprintf(keybuf.data(), keybuf.size(), key, arch); - return keybuf.data(); - }; - - gguf_set_val_str(fctx, LLM_KV_GENERAL_ARCHITECTURE, arch); - gguf_set_val_u32(fctx, LLM_KV_GENERAL_FILE_TYPE, ftype); - - gguf_set_val_u32(fctx, kv(LLM_KV_CONTEXT_LENGTH), model->hparams.n_ctx); - gguf_set_val_u32(fctx, kv(LLM_KV_EMBEDDING_LENGTH), model->hparams.n_embd); - gguf_set_val_u32(fctx, kv(LLM_KV_FEED_FORWARD_LENGTH), model->hparams.n_ff); - gguf_set_val_u32(fctx, kv(LLM_KV_ATTENTION_HEAD_COUNT), model->hparams.n_head); - gguf_set_val_u32(fctx, kv(LLM_KV_ATTENTION_HEAD_COUNT_KV), model->hparams.n_head_kv); - gguf_set_val_u32(fctx, kv(LLM_KV_BLOCK_COUNT), model->hparams.n_layer); - gguf_set_val_u32(fctx, kv(LLM_KV_ROPE_DIMENSION_COUNT), model->hparams.n_embd_head()); - gguf_set_val_f32(fctx, kv(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS), model->hparams.f_norm_rms_eps); - gguf_set_val_f32(fctx, kv(LLM_KV_ROPE_FREQ_BASE), model->hparams.rope_freq_base); - gguf_set_val_f32(fctx, kv(LLM_KV_ROPE_SCALE_LINEAR), model->hparams.rope_freq_scale); - - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_TOKEN_EMBD, lora->hparams.n_rank_tok_embeddings); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_OUTPUT_NORM, lora->hparams.n_rank_norm); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_OUTPUT, lora->hparams.n_rank_output); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_ATTN_NORM, lora->hparams.n_rank_attention_norm); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_ATTN_Q, lora->hparams.n_rank_wq); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_ATTN_K, lora->hparams.n_rank_wk); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_ATTN_V, lora->hparams.n_rank_wv); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_ATTN_OUT, lora->hparams.n_rank_wo); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_FFN_NORM, lora->hparams.n_rank_ffn_norm); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_FFN_GATE, lora->hparams.n_rank_ffn_gate); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_FFN_DOWN, lora->hparams.n_rank_ffn_down); - gguf_set_val_u32(fctx, LLM_KV_TRAINING_LORA_RANK_FFN_UP, lora->hparams.n_rank_ffn_up); - - gguf_add_tensor(fctx, lora->tok_embeddings_a); - gguf_add_tensor(fctx, lora->tok_embeddings_b); - gguf_add_tensor(fctx, lora->norm_a); - gguf_add_tensor(fctx, lora->norm_b); - gguf_add_tensor(fctx, lora->output_a); - gguf_add_tensor(fctx, lora->output_b); - - for (uint32_t i = 0; i < lora->layers.size(); ++i) { - auto & layer = lora->layers[i]; - - gguf_add_tensor(fctx, layer.attention_norm_a); - gguf_add_tensor(fctx, layer.attention_norm_b); - gguf_add_tensor(fctx, layer.wq_a); - gguf_add_tensor(fctx, layer.wq_b); - gguf_add_tensor(fctx, layer.wk_a); - gguf_add_tensor(fctx, layer.wk_b); - gguf_add_tensor(fctx, layer.wv_a); - gguf_add_tensor(fctx, layer.wv_b); - gguf_add_tensor(fctx, layer.wo_a); - gguf_add_tensor(fctx, layer.wo_b); - gguf_add_tensor(fctx, layer.ffn_norm_a); - gguf_add_tensor(fctx, layer.ffn_norm_b); - gguf_add_tensor(fctx, layer.ffn_gate_a); - gguf_add_tensor(fctx, layer.ffn_gate_b); - gguf_add_tensor(fctx, layer.ffn_down_a); - gguf_add_tensor(fctx, layer.ffn_down_b); - gguf_add_tensor(fctx, layer.ffn_up_a); - gguf_add_tensor(fctx, layer.ffn_up_b); - } -} - -static void load_checkpoint_lora_gguf(struct gguf_context * fctx, struct ggml_context * f_ggml_ctx, struct my_llama_model * model, struct my_llama_lora * lora, struct train_state * train) { - std::string train_type = LLM_KV_TRAINING_TYPE_FINETUNE_LORA; - GGUF_GET_KEY(fctx, train_type, gguf_get_val_str, GGUF_TYPE_STRING, false, LLM_KV_TRAINING_TYPE); - GGML_ASSERT(train_type == LLM_KV_TRAINING_TYPE_FINETUNE_LORA); - - load_train_state_gguf(fctx, f_ggml_ctx, train); - load_llama_lora_gguf(fctx, f_ggml_ctx, model, lora); -} - -static void save_checkpoint_lora_gguf(struct gguf_context * fctx, struct my_llama_model * model, struct my_llama_lora * lora, struct train_state * train) { - gguf_set_val_str(fctx, LLM_KV_TRAINING_TYPE, LLM_KV_TRAINING_TYPE_FINETUNE_LORA); - save_llama_lora_gguf(fctx, model, lora); - save_train_state_gguf(fctx, train); -} - -static bool load_checkpoint_lora_file(const char * filename, struct my_llama_model * model, struct my_llama_lora * lora, struct train_state * train) { - struct ggml_context * f_ggml_ctx; - struct gguf_init_params params; - params.no_alloc = false; - params.ctx = &f_ggml_ctx; - struct gguf_context * fctx = gguf_init_from_file(filename, params); - if (fctx == NULL) { - return false; - } - - load_checkpoint_lora_gguf(fctx, f_ggml_ctx, model, lora, train); - - gguf_free(fctx); - return true; -} - -static void save_checkpoint_lora_file(const char * filename, struct my_llama_model * model, struct my_llama_lora * lora, struct train_state * train) { - printf("%s: saving to %s\n", __func__, filename); - struct gguf_context * fctx = gguf_init_empty(); - - save_checkpoint_lora_gguf(fctx, model, lora, train); - - // write file - const bool only_meta = false; - gguf_write_to_file(fctx, filename, only_meta); - gguf_free(fctx); -} - -struct llama_file { - // use FILE * so we don't have to re-open the file to mmap - FILE * fp; - size_t size; - - llama_file(const char * fname, const char * mode) { - fp = std::fopen(fname, mode); - if (fp == NULL) { - size = 0; - } else { - seek(0, SEEK_END); - size = tell(); - seek(0, SEEK_SET); - } - } - - size_t tell() const { -#ifdef _WIN32 - __int64 ret = _ftelli64(fp); -#else - long ret = std::ftell(fp); -#endif - GGML_ASSERT(ret != -1); // this really shouldn't fail - return (size_t) ret; - } - - void seek(size_t offset, int whence) { -#ifdef _WIN32 - int ret = _fseeki64(fp, (__int64) offset, whence); -#else - int ret = std::fseek(fp, (long) offset, whence); -#endif - GGML_ASSERT(ret == 0); // same - } - - void read_raw(void * ptr, size_t size) { - if (size == 0) { - return; - } - errno = 0; - std::size_t ret = std::fread(ptr, size, 1, fp); - if (ferror(fp)) { - die_fmt("read error: %s", strerror(errno)); - } - if (ret != 1) { - die("unexpectedly reached end of file"); - } - } - - std::uint32_t read_u32() { - std::uint32_t ret; - read_raw(&ret, sizeof(ret)); - return ret; - } - - std::string read_string(std::uint32_t len) { - std::vector chars(len); - read_raw(chars.data(), len); - return std::string(chars.data(), len); - } - - void write_raw(const void * ptr, size_t size) { - if (size == 0) { - return; - } - errno = 0; - size_t ret = std::fwrite(ptr, size, 1, fp); - if (ret != 1) { - die_fmt("write error: %s", strerror(errno)); - } - } - - void write_u32(std::uint32_t val) { - write_raw(&val, sizeof(val)); - } - - ~llama_file() { - if (fp) { - std::fclose(fp); - } - } -}; - -static void write_tensor(struct llama_file * file, struct ggml_tensor * tensor, const char * name) { - if (tensor == NULL) { - file->write_u32(0); - file->write_u32(0); - file->write_u32(GGML_TYPE_F32); - file->seek((0-file->tell()) & 31, SEEK_CUR); - return; - } - if (name == NULL) { - name = ggml_get_name(tensor); - } - uint32_t name_len = strlen(name); - uint32_t nd = ggml_n_dims(tensor); - uint32_t ne[4] = { (uint32_t)tensor->ne[0], - (uint32_t)tensor->ne[1], - (uint32_t)tensor->ne[2], - (uint32_t)tensor->ne[3] }; - file->write_u32(nd); - file->write_u32(name_len); - file->write_u32(tensor->type); - file->write_raw(ne, sizeof(ne[0]) * nd); - file->write_raw(name, name_len); - file->seek((0-file->tell()) & 31, SEEK_CUR); - file->write_raw(tensor->data, ggml_nbytes(tensor)); -} - -static void save_as_llama_lora(const char * filename, struct my_llama_lora * lora) { - printf("%s: saving to %s\n", __func__, filename); - struct llama_file file(filename, "wb"); - if (file.fp == NULL) { - return; - } - - std::vector tn_buf; - tn_buf.resize(GGML_MAX_NAME); - - auto tn = [&tn_buf](const char * key, const char * suffix) -> const char * { - snprintf(tn_buf.data(), tn_buf.size(), "%s%s", key, suffix); - return tn_buf.data(); - }; - - auto tni = [&tn_buf](const char * key, int bid, const char * suffix) -> const char * { - snprintf(tn_buf.data(), tn_buf.size(), key, bid); - std::string s = tn_buf.data(); - snprintf(tn_buf.data(), tn_buf.size(), "%s%s", s.c_str(), suffix); - return tn_buf.data(); - }; - - // write_magic - file.write_u32(LLAMA_FILE_MAGIC_GGLA); // magic - file.write_u32(1); // version - // write_hparams - file.write_u32(lora->hparams.lora_r); - file.write_u32(lora->hparams.lora_alpha); - // write tensors - write_tensor(&file, lora->tok_embeddings_a, tn(LLM_TENSOR_TOKEN_EMBD, ".weight.loraA")); - write_tensor(&file, lora->tok_embeddings_b, tn(LLM_TENSOR_TOKEN_EMBD, ".weight.loraB")); - write_tensor(&file, lora->norm_a, tn(LLM_TENSOR_OUTPUT_NORM, ".weight.loraA")); - write_tensor(&file, lora->norm_b, tn(LLM_TENSOR_OUTPUT_NORM, ".weight.loraB")); - write_tensor(&file, lora->output_a, tn(LLM_TENSOR_OUTPUT, ".weight.loraA")); - write_tensor(&file, lora->output_b, tn(LLM_TENSOR_OUTPUT, ".weight.loraB")); - for (uint32_t i = 0; i < lora->layers.size(); ++i) { - auto & layer = lora->layers[i]; - write_tensor(&file, layer.attention_norm_a, tni(LLM_TENSOR_ATTN_NORM, i, ".weight.loraA")); - write_tensor(&file, layer.attention_norm_b, tni(LLM_TENSOR_ATTN_NORM, i, ".weight.loraB")); - write_tensor(&file, layer.wq_a, tni(LLM_TENSOR_ATTN_Q, i, ".weight.loraA")); - write_tensor(&file, layer.wq_b, tni(LLM_TENSOR_ATTN_Q, i, ".weight.loraB")); - write_tensor(&file, layer.wk_a, tni(LLM_TENSOR_ATTN_K, i, ".weight.loraA")); - write_tensor(&file, layer.wk_b, tni(LLM_TENSOR_ATTN_K, i, ".weight.loraB")); - write_tensor(&file, layer.wv_a, tni(LLM_TENSOR_ATTN_V, i, ".weight.loraA")); - write_tensor(&file, layer.wv_b, tni(LLM_TENSOR_ATTN_V, i, ".weight.loraB")); - write_tensor(&file, layer.wo_a, tni(LLM_TENSOR_ATTN_OUT, i, ".weight.loraA")); - write_tensor(&file, layer.wo_b, tni(LLM_TENSOR_ATTN_OUT, i, ".weight.loraB")); - write_tensor(&file, layer.ffn_norm_a, tni(LLM_TENSOR_FFN_NORM, i, ".weight.loraA")); - write_tensor(&file, layer.ffn_norm_b, tni(LLM_TENSOR_FFN_NORM, i, ".weight.loraB")); - write_tensor(&file, layer.ffn_gate_a, tni(LLM_TENSOR_FFN_GATE, i, ".weight.loraA")); - write_tensor(&file, layer.ffn_gate_b, tni(LLM_TENSOR_FFN_GATE, i, ".weight.loraB")); - write_tensor(&file, layer.ffn_down_a, tni(LLM_TENSOR_FFN_DOWN, i, ".weight.loraA")); - write_tensor(&file, layer.ffn_down_b, tni(LLM_TENSOR_FFN_DOWN, i, ".weight.loraB")); - write_tensor(&file, layer.ffn_up_a, tni(LLM_TENSOR_FFN_UP, i, ".weight.loraA")); - write_tensor(&file, layer.ffn_up_b, tni(LLM_TENSOR_FFN_UP, i, ".weight.loraB")); - } -} - -struct train_params { - struct train_params_common common; - - const char * fn_model_base; - const char * fn_lora_out; - - bool only_write_lora; - - float f_norm_rms_eps; - float rope_freq_base; - float rope_freq_scale; - - bool custom_f_norm_rms_eps; - bool custom_rope_freq_base; - bool custom_rope_freq_scale; - - int32_t lora_r; - int32_t lora_alpha; - bool custom_lora_alpha; - - uint32_t n_rank_attention_norm; - uint32_t n_rank_wq; - uint32_t n_rank_wk; - uint32_t n_rank_wv; - uint32_t n_rank_wo; - uint32_t n_rank_ffn_norm; - uint32_t n_rank_ffn_gate; - uint32_t n_rank_ffn_down; - uint32_t n_rank_ffn_up; - uint32_t n_rank_tok_embeddings; - uint32_t n_rank_norm; - uint32_t n_rank_output; - - bool custom_n_rank_attention_norm; - bool custom_n_rank_wq; - bool custom_n_rank_wk; - bool custom_n_rank_wv; - bool custom_n_rank_wo; - bool custom_n_rank_ffn_norm; - bool custom_n_rank_ffn_gate; - bool custom_n_rank_ffn_down; - bool custom_n_rank_ffn_up; - bool custom_n_rank_tok_embeddings; - bool custom_n_rank_norm; - bool custom_n_rank_output; -}; - -static struct train_params get_default_train_params() { - struct train_params params; - params.common = get_default_train_params_common(); - params.fn_model_base = ""; - params.fn_lora_out = "ggml-lora-ITERATION-f32.gguf"; - - params.only_write_lora = false; - - params.f_norm_rms_eps = 1e-5f; - params.rope_freq_base = 10000.0f; - params.rope_freq_scale = 1.0f; - - params.custom_f_norm_rms_eps = false; - params.custom_rope_freq_base = false; - params.custom_rope_freq_scale = false; - - params.lora_r = 4; - params.lora_alpha = 4; - params.custom_lora_alpha = false; - - params.n_rank_attention_norm = 1; - params.n_rank_wq = 4; - params.n_rank_wk = 4; - params.n_rank_wv = 4; - params.n_rank_wo = 4; - params.n_rank_ffn_norm = 1; - params.n_rank_ffn_gate = 4; - params.n_rank_ffn_down = 4; - params.n_rank_ffn_up = 4; - params.n_rank_tok_embeddings = 4; - params.n_rank_norm = 1; - params.n_rank_output = 4; - - params.custom_n_rank_attention_norm = false; - params.custom_n_rank_wq = false; - params.custom_n_rank_wk = false; - params.custom_n_rank_wv = false; - params.custom_n_rank_wo = false; - params.custom_n_rank_ffn_norm = false; - params.custom_n_rank_ffn_gate = false; - params.custom_n_rank_ffn_down = false; - params.custom_n_rank_ffn_up = false; - params.custom_n_rank_tok_embeddings = false; - params.custom_n_rank_norm = false; - params.custom_n_rank_output = false; - - return params; -} - -static void train_print_usage(int argc, char ** argv, const struct train_params * params) { - fprintf(stderr, "usage: %s [options]\n", argv[0]); - fprintf(stderr, "\n"); - fprintf(stderr, "options:\n"); - fprintf(stderr, " -h, --help show this help message and exit\n"); - - fprintf(stderr, " --model-base FNAME model path from which to load base model (default '%s')\n", params->fn_model_base); - fprintf(stderr, " --lora-out FNAME path to save llama lora (default '%s')\n", params->fn_lora_out); - fprintf(stderr, " --only-write-lora only save llama lora, don't do any training. use this if you only want to convert a checkpoint to a lora adapter.\n"); - fprintf(stderr, " --norm-rms-eps F RMS-Norm epsilon value (default %f)\n", params->f_norm_rms_eps); - fprintf(stderr, " --rope-freq-base F Frequency base for ROPE (default %f)\n", params->rope_freq_base); - fprintf(stderr, " --rope-freq-scale F Frequency scale for ROPE (default %f)\n", params->rope_freq_scale); - fprintf(stderr, " --lora-alpha N LORA alpha : resulting LORA scaling is alpha/r. (default %d)\n", params->lora_alpha); - fprintf(stderr, " --lora-r N LORA r: default rank. Also specifies resulting scaling together with lora-alpha. (default %d)\n", params->lora_r); - fprintf(stderr, " --rank-att-norm N LORA rank for attention norm tensor, overrides default rank. Norm tensors should generally have rank 1.\n"); - fprintf(stderr, " --rank-ffn-norm N LORA rank for feed-forward norm tensor, overrides default rank. Norm tensors should generally have rank 1.\n"); - fprintf(stderr, " --rank-out-norm N LORA rank for output norm tensor, overrides default rank. Norm tensors should generally have rank 1.\n"); - fprintf(stderr, " --rank-tok-embd N LORA rank for token embeddings tensor, overrides default rank.\n"); - fprintf(stderr, " --rank-out N LORA rank for output tensor, overrides default rank.\n"); - fprintf(stderr, " --rank-wq N LORA rank for wq tensor, overrides default rank.\n"); - fprintf(stderr, " --rank-wk N LORA rank for wk tensor, overrides default rank.\n"); - fprintf(stderr, " --rank-wv N LORA rank for wv tensor, overrides default rank.\n"); - fprintf(stderr, " --rank-wo N LORA rank for wo tensor, overrides default rank.\n"); - fprintf(stderr, " --rank-ffn_gate N LORA rank for ffn_gate tensor, overrides default rank.\n"); - fprintf(stderr, " --rank-ffn_down N LORA rank for ffn_down tensor, overrides default rank.\n"); - fprintf(stderr, " --rank-ffn_up N LORA rank for ffn_up tensor, overrides default rank.\n"); - - print_common_train_usage(argc, argv, ¶ms->common); -} - -static bool train_params_parse(int argc, char ** argv, struct train_params * params) { - bool invalid_param = false; - std::string arg; - struct train_params default_params = get_default_train_params(); - const std::string arg_prefix = "--"; - - for (int i = 1; i < argc; i++) { - arg = argv[i]; - if (arg.compare(0, arg_prefix.size(), arg_prefix) == 0) { - std::replace(arg.begin(), arg.end(), '_', '-'); - } - - if (consume_common_train_arg(argc, argv, &i, ¶ms->common, &invalid_param)) { - if (invalid_param) { - break; - } else if (params->common.print_usage) { - train_print_usage(argc, argv, &default_params); - exit(0); - } - } else if (arg == "--model-base") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->fn_model_base = argv[i]; - } else if (arg == "--lora-out") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->fn_lora_out = argv[i]; - } else if (arg == "--only-write-lora") { - params->only_write_lora = true; - } else if (arg == "--norm-rms-eps") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->f_norm_rms_eps = std::stof(argv[i]); - params->custom_f_norm_rms_eps = true; - } else if (arg == "--rope-freq-base") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->rope_freq_base = std::stof(argv[i]); - params->custom_rope_freq_base = true; - } else if (arg == "--rope-freq-scale") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->rope_freq_scale = std::stof(argv[i]); - params->custom_rope_freq_scale = true; - } else if (arg == "--lora-alpha") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->lora_alpha = std::stoi(argv[i]); - params->custom_lora_alpha = true; - } else if (arg == "--lora-r") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->lora_r = std::stoi(argv[i]); - } else if (arg == "--rank-att-norm") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_attention_norm = std::stoi(argv[i]); - params->custom_n_rank_attention_norm = true; - } else if (arg == "--rank-ffn-norm") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_ffn_norm = std::stoi(argv[i]); - params->custom_n_rank_ffn_norm = true; - } else if (arg == "--rank-out-norm") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_norm = std::stoi(argv[i]); - params->custom_n_rank_norm = true; - } else if (arg == "--rank-tok-embd") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_tok_embeddings = std::stoi(argv[i]); - params->custom_n_rank_tok_embeddings = true; - } else if (arg == "--rank-out") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_output = std::stoi(argv[i]); - params->custom_n_rank_output = true; - } else if (arg == "--rank-wq") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_wq = std::stoi(argv[i]); - params->custom_n_rank_wq = true; - } else if (arg == "--rank-wk") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_wk = std::stoi(argv[i]); - params->custom_n_rank_wk = true; - } else if (arg == "--rank-wv") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_wv = std::stoi(argv[i]); - params->custom_n_rank_wv = true; - } else if (arg == "--rank-wo") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_wo = std::stoi(argv[i]); - params->custom_n_rank_wo = true; - } else if (arg == "--rank-ffn_gate") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_ffn_gate = std::stoi(argv[i]); - params->custom_n_rank_ffn_gate = true; - } else if (arg == "--rank-ffn_down") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_ffn_down = std::stoi(argv[i]); - params->custom_n_rank_ffn_down = true; - } else if (arg == "--rank-ffn_up") { - if (++i >= argc) { - invalid_param = true; - break; - } - params->n_rank_ffn_up = std::stoi(argv[i]); - params->custom_n_rank_ffn_up = true; - } else { - fprintf(stderr, "error: unknown argument: %s\n", arg.c_str()); - train_print_usage(argc, argv, &default_params); - exit(1); - } - } - if (invalid_param) { - fprintf(stderr, "error: invalid parameter for argument: %s\n", arg.c_str()); - train_print_usage(argc, argv, &default_params); - exit(1); - } - finish_processing_train_args(¶ms->common); - return true; -} - -struct save_train_files_data { - const char * fn_checkpoint_out; - const char * fn_lora_out; - const char * pattern_fn_it; - const char * fn_latest; - struct my_llama_model * model; - struct my_llama_lora * lora; -}; - -static void save_train_files(void * vdata, struct train_state * train) { - struct save_train_files_data * data = (struct save_train_files_data *) vdata; - - int64_t iter = train->opt->iter; - - if (strlen(data->fn_checkpoint_out) > 0) { - save_checkpoint_lora_file(get_train_filename(data->fn_checkpoint_out, data->pattern_fn_it, data->fn_latest, iter).c_str(), data->model, data->lora, train); - save_checkpoint_lora_file(get_train_filename(data->fn_checkpoint_out, data->pattern_fn_it, data->fn_latest, -1 ).c_str(), data->model, data->lora, train); - } - if (strlen(data->fn_lora_out) > 0) { - save_as_llama_lora(get_train_filename(data->fn_lora_out, data->pattern_fn_it, data->fn_latest, iter).c_str(), data->lora); - save_as_llama_lora(get_train_filename(data->fn_lora_out, data->pattern_fn_it, data->fn_latest, -1 ).c_str(), data->lora); - } -} - -static int64_t get_parameter_count(struct my_llama_lora* lora) { - int64_t nx = 0; - nx += ggml_nelements(lora->tok_embeddings_a); - nx += ggml_nelements(lora->tok_embeddings_b); - nx += ggml_nelements(lora->norm_a); - nx += ggml_nelements(lora->norm_b); - nx += ggml_nelements(lora->output_a); - nx += ggml_nelements(lora->output_b); - - for (uint32_t i = 0; i < lora->layers.size(); ++i) { - auto & layer = lora->layers[i]; - nx += ggml_nelements(layer.attention_norm_a); - nx += ggml_nelements(layer.attention_norm_b); - nx += ggml_nelements(layer.wq_a); - nx += ggml_nelements(layer.wq_b); - nx += ggml_nelements(layer.wk_a); - nx += ggml_nelements(layer.wk_b); - nx += ggml_nelements(layer.wv_a); - nx += ggml_nelements(layer.wv_b); - nx += ggml_nelements(layer.wo_a); - nx += ggml_nelements(layer.wo_b); - nx += ggml_nelements(layer.ffn_norm_a); - nx += ggml_nelements(layer.ffn_norm_b); - nx += ggml_nelements(layer.ffn_gate_a); - nx += ggml_nelements(layer.ffn_gate_b); - nx += ggml_nelements(layer.ffn_down_a); - nx += ggml_nelements(layer.ffn_down_b); - nx += ggml_nelements(layer.ffn_up_a); - nx += ggml_nelements(layer.ffn_up_b); - } - return nx; -} - -int main(int argc, char ** argv) { - struct train_params params = get_default_train_params(); - - if (!train_params_parse(argc, argv, ¶ms)) { - return 1; - } - - if (params.common.seed == LLAMA_DEFAULT_SEED) { - params.common.seed = time(NULL); - } - printf("%s: seed: %u\n", __func__, params.common.seed); - srand(params.common.seed); - - struct llama_model_params llama_mparams = llama_model_default_params(); - llama_mparams.n_gpu_layers = params.common.n_gpu_layers; - llama_mparams.vocab_only = false; - - printf("%s: model base = '%s'\n", __func__, params.fn_model_base); - struct llama_model * lmodel = llama_load_model_from_file(params.fn_model_base, llama_mparams); - - struct llama_context_params llama_cparams = llama_context_default_params(); - struct llama_context * lctx = llama_new_context_with_model(lmodel, llama_cparams); - - struct my_llama_model model; - init_model(lmodel, &model, params.fn_model_base, params.common.n_ctx); - - struct my_llama_lora lora; - - struct train_state * train = init_train_state(); - struct ggml_opt_context * opt = train->opt; - - // set params from command line - if (params.custom_f_norm_rms_eps) { - model.hparams.f_norm_rms_eps = params.f_norm_rms_eps; - } - if (params.custom_rope_freq_base) { - model.hparams.rope_freq_base = params.rope_freq_base; - } - if (params.custom_rope_freq_scale) { - model.hparams.rope_freq_scale = params.rope_freq_scale; - } - lora.hparams.lora_r = params.lora_r; - lora.hparams.lora_alpha = params.custom_lora_alpha ? params.lora_alpha : params.lora_r; - uint32_t n_rank_attention_norm = params.custom_n_rank_attention_norm ? params.n_rank_attention_norm : 1; - uint32_t n_rank_wq = params.custom_n_rank_wq ? params.n_rank_wq : params.lora_r; - uint32_t n_rank_wk = params.custom_n_rank_wk ? params.n_rank_wk : params.lora_r; - uint32_t n_rank_wv = params.custom_n_rank_wv ? params.n_rank_wv : params.lora_r; - uint32_t n_rank_wo = params.custom_n_rank_wo ? params.n_rank_wo : params.lora_r; - uint32_t n_rank_ffn_norm = params.custom_n_rank_ffn_norm ? params.n_rank_ffn_norm : 1; - uint32_t n_rank_ffn_gate = params.custom_n_rank_ffn_gate ? params.n_rank_ffn_gate : params.lora_r; - uint32_t n_rank_ffn_down = params.custom_n_rank_ffn_down ? params.n_rank_ffn_down : params.lora_r; - uint32_t n_rank_ffn_up = params.custom_n_rank_ffn_up ? params.n_rank_ffn_up : params.lora_r; - uint32_t n_rank_tok_embeddings = params.custom_n_rank_tok_embeddings ? params.n_rank_tok_embeddings : params.lora_r; - uint32_t n_rank_norm = params.custom_n_rank_norm ? params.n_rank_norm : 1; - uint32_t n_rank_output = params.custom_n_rank_output ? params.n_rank_output : params.lora_r; - lora.hparams.n_rank_attention_norm = n_rank_attention_norm; - lora.hparams.n_rank_wq = n_rank_wq; - lora.hparams.n_rank_wk = n_rank_wk; - lora.hparams.n_rank_wv = n_rank_wv; - lora.hparams.n_rank_wo = n_rank_wo; - lora.hparams.n_rank_ffn_norm = n_rank_ffn_norm; - lora.hparams.n_rank_ffn_gate = n_rank_ffn_gate; - lora.hparams.n_rank_ffn_down = n_rank_ffn_down; - lora.hparams.n_rank_ffn_up = n_rank_ffn_up; - lora.hparams.n_rank_tok_embeddings = n_rank_tok_embeddings; - lora.hparams.n_rank_norm = n_rank_norm; - lora.hparams.n_rank_output = n_rank_output; - - // set opt params from command line - opt->params = ggml_opt_default_params(GGML_OPT_TYPE_ADAM); - opt->params.print_forward_graph = false; - opt->params.print_backward_graph = false; - opt->params.graph_size = LLAMA_TRAIN_MAX_NODES; - opt->params.n_threads = params.common.n_threads; - opt->params.past = params.common.opt_past; - opt->params.delta = params.common.opt_delta; - opt->params.max_no_improvement = params.common.opt_max_no_improvement; - opt->params.n_gradient_accumulation = params.common.n_gradient_accumulation; - opt->params.adam.n_iter = params.common.adam_n_iter; - opt->params.adam.sched = 1.0f; - opt->params.adam.alpha = params.common.adam_alpha; - opt->params.adam.decay = params.common.adam_decay; - opt->params.adam.decay_min_ndim = params.common.adam_decay_min_ndim; - opt->params.adam.beta1 = params.common.adam_beta1; - opt->params.adam.beta2 = params.common.adam_beta2; - opt->params.adam.gclip = params.common.adam_gclip; - opt->params.adam.eps_f = params.common.adam_eps_f; - - printf("%s: init model\n", __func__); - bool existed = load_checkpoint_lora_file(params.common.fn_checkpoint_in, &model, &lora, train); - - if (existed) { - // overwrite last n_ctx with user provided n_ctx - if (params.common.custom_n_ctx) { - model.hparams.n_ctx = params.common.n_ctx; - } - - const bool opt_param_count_changed = ( - (lora.hparams.n_rank_attention_norm != n_rank_attention_norm) - || (lora.hparams.n_rank_wq != n_rank_wq) - || (lora.hparams.n_rank_wk != n_rank_wk) - || (lora.hparams.n_rank_wv != n_rank_wv) - || (lora.hparams.n_rank_wo != n_rank_wo) - || (lora.hparams.n_rank_ffn_norm != n_rank_ffn_norm) - || (lora.hparams.n_rank_ffn_gate != n_rank_ffn_gate) - || (lora.hparams.n_rank_ffn_down != n_rank_ffn_down) - || (lora.hparams.n_rank_ffn_up != n_rank_ffn_up) - || (lora.hparams.n_rank_tok_embeddings != n_rank_tok_embeddings) - || (lora.hparams.n_rank_norm != n_rank_norm) - || (lora.hparams.n_rank_output != n_rank_output) - ); - - const bool opt_past_changed = opt->params.past != params.common.opt_past; - - if (opt_param_count_changed) { - print_lora_params(&lora.hparams); - die("Provided rank differs from checkpoint file. To use different rank start finetune from scratch with empty input checkpoint, e.g --checkpoint-in ''. Aborting."); - // need to discard previous optimizer gradient statistics and opt_init with new shapes - // TODO - } - if (opt_past_changed) { - die("Optimizer parameter '--opt-past N' differs from checkpoint file. To use different value finetune from scratch with empty input checkpoint, e.g --checkpoint-in ''. Aborting"); - // need to discard previous optimizer past function value statistics and opt_init with new shapes - // TODO - } - } else { // existed == false - init_lora(&model, &lora); - randomize_lora(&lora, params.common.seed, 0.0f, 1.0f, -1.0f, +1.0f); - if (!params.only_write_lora) { - ggml_opt_init(opt->ctx, opt, opt->params, get_parameter_count(&lora)); - } - } - opt->iter = train->train_its; - - print_params(&model.hparams); - print_lora_params(&lora.hparams); - printf("%s: total train_iterations %llu\n", __func__, (long long unsigned) train->train_its); - printf("%s: seen train_samples %llu\n", __func__, (long long unsigned) train->train_samples); - printf("%s: seen train_tokens %llu\n", __func__, (long long unsigned) train->train_tokens); - printf("%s: completed train_epochs %llu\n", __func__, (long long unsigned) train->train_epochs); - printf("%s: lora_size = %zu bytes (%.1f MB)\n", __func__, (ggml_used_mem(lora.ctx) + ggml_backend_buffer_get_size(lora.data)), (float) (ggml_used_mem(lora.ctx) + ggml_backend_buffer_get_size(lora.data)) / (1024.0f*1024.0f)); - - if (params.only_write_lora) { - save_train_files_data save_data; - save_data.fn_checkpoint_out = ""; - save_data.fn_lora_out = params.fn_lora_out; - save_data.pattern_fn_it = params.common.pattern_fn_it; - save_data.fn_latest = params.common.fn_latest; - save_data.model = &model; - save_data.lora = &lora; - - save_train_files(&save_data, train); - - free_train_state(train); - ggml_free(lora.ctx); - llama_free(lctx); - llama_free_model(lmodel); - return 0; - } - - printf("%s: opt_size = %zu bytes (%.1f MB)\n", __func__, ggml_get_mem_size(opt->ctx), (float) ggml_get_mem_size(opt->ctx) / (1024.0f*1024.0f)); - printf("%s: opt iter %d\n", __func__, opt->iter); - - int n_tokens = model.hparams.n_ctx; - int n_vocab = model.hparams.n_vocab; - int n_batch = params.common.n_batch; - - // context for input tensors without their data - struct ggml_init_params ctx_input_params = { - ggml_tensor_overhead() * 2, // mem_size - NULL, // mem_buffer - true, // no_alloc - }; - struct ggml_context * ctx_input = ggml_init(ctx_input_params); - - // the input tensors - struct ggml_tensor * tokens_input = ggml_new_tensor_2d(ctx_input, GGML_TYPE_I32, n_tokens, n_batch); - struct ggml_tensor * target_probs = ggml_new_tensor_3d(ctx_input, GGML_TYPE_F32, n_vocab, n_tokens, n_batch); - - // allocate input tensors - // measure required memory for input tensors - ggml_backend_buffer_t input_data = ggml_backend_alloc_ctx_tensors_from_buft(ctx_input, ggml_backend_cpu_buffer_type()); - size_t max_input_size = ggml_backend_buffer_get_size(input_data); - printf("%s: input_size = %zu bytes (%.1f MB)\n", __func__, max_input_size, (float) max_input_size / (1024.0f*1024.0f)); - - // context for compute tensors without their data - const size_t estimated_compute_size_wo_data = ( - 2*LLAMA_TRAIN_MAX_NODES*ggml_tensor_overhead() + - (params.common.use_checkpointing ? 3 : 2)*(GGML_OBJECT_SIZE+ggml_graph_overhead_custom(LLAMA_TRAIN_MAX_NODES, true)) - ); - struct ggml_init_params ctx_compute_params = { - estimated_compute_size_wo_data, // mem_size - NULL, // mem_buffer - true, // no_alloc - }; - struct ggml_context * ctx_compute = NULL; - - struct ggml_tensor * loss = NULL; - struct ggml_tensor * logits = NULL; - - struct ggml_cgraph * gf = NULL; - struct ggml_cgraph * gb = NULL; - struct ggml_cgraph * gb_tmp = NULL; - - // measure required memory for compute tensors - size_t best_compute_size = SIZE_MAX; - enum ggml_cgraph_eval_order best_order = GGML_CGRAPH_EVAL_ORDER_COUNT; - // find best evaluation order - for (unsigned order = 0; order < (unsigned) GGML_CGRAPH_EVAL_ORDER_COUNT; ++order) { - ctx_compute = ggml_init(ctx_compute_params); - ggml_gallocr_t alloc = ggml_gallocr_new(ggml_backend_cpu_buffer_type()); - gf = ggml_new_graph_custom(ctx_compute, LLAMA_TRAIN_MAX_NODES, true); - gf->order = (enum ggml_cgraph_eval_order) order; - gb = ggml_new_graph_custom(ctx_compute, LLAMA_TRAIN_MAX_NODES, true); - gb_tmp = params.common.use_checkpointing - ? ggml_new_graph_custom(ctx_compute, LLAMA_TRAIN_MAX_NODES, true) - : NULL; - loss = llama_build_lora_finetune_graphs( - &model, &lora, alloc, ctx_compute, - gf, gb, gb_tmp, - &logits, tokens_input, target_probs, - n_tokens, n_batch, - params.common.use_flash, - params.common.use_checkpointing, - true - ); - size_t max_compute_size = ggml_gallocr_get_buffer_size(alloc, 0); // FIXME: this will still allocate the buffer - if (max_compute_size < best_compute_size) { - best_compute_size = max_compute_size; - best_order = gf->order; - } - ggml_gallocr_free(alloc); - ggml_free(ctx_compute); - } - size_t max_compute_size = best_compute_size; - printf("%s: compute_size = %zu bytes (%.1f MB)\n", __func__, max_compute_size, (float) max_compute_size / (1024.0f*1024.0f)); - printf("%s: evaluation order = %s\n", __func__, - (best_order == GGML_CGRAPH_EVAL_ORDER_LEFT_TO_RIGHT) ? "LEFT_TO_RIGHT" : - (best_order == GGML_CGRAPH_EVAL_ORDER_RIGHT_TO_LEFT) ? "RIGHT_TO_LEFT" : - "invalid"); - - // allocate compute tensors - ctx_compute = ggml_init(ctx_compute_params); - ggml_gallocr_t alloc = ggml_gallocr_new(ggml_backend_cpu_buffer_type()); - gf = ggml_new_graph_custom(ctx_compute, LLAMA_TRAIN_MAX_NODES, true); - gf->order = best_order; - gb = ggml_new_graph_custom(ctx_compute, LLAMA_TRAIN_MAX_NODES, true); - gb_tmp = params.common.use_checkpointing - ? ggml_new_graph_custom(ctx_compute, LLAMA_TRAIN_MAX_NODES, true) - : NULL; - loss = llama_build_lora_finetune_graphs( - &model, &lora, alloc, ctx_compute, - gf, gb, gb_tmp, - &logits, tokens_input, target_probs, - n_tokens, n_batch, - params.common.use_flash, - params.common.use_checkpointing, - false - ); - - // tokenize data - std::vector train_tokens; - std::vector train_samples_begin; - std::vector train_samples_size; - printf("%s: tokenize training data from %s\n", __func__, params.common.fn_train_data); - printf("%s: sample-start: %s\n", __func__, params.common.sample_start.c_str()); - printf("%s: include-sample-start: %s\n", __func__, params.common.include_sample_start ? "true" : "false"); - tokenize_file(lctx, - params.common.fn_train_data, - params.common.sample_start, - params.common.include_sample_start, - params.common.overlapping_samples, - n_tokens, - train_tokens, - train_samples_begin, - train_samples_size); - GGML_ASSERT(train_samples_begin.size() == train_samples_size.size()); - - printf("%s: number of training tokens: %zu\n", __func__, train_tokens.size()); - - std::vector token_noccurs; - token_noccurs.resize(model.hparams.n_vocab, 0); - for (unsigned int i = 0; i < train_tokens.size(); ++i) { - ++token_noccurs[train_tokens[i]]; - } - int n_unique_tokens = 0; - for (unsigned int i = 0; i < token_noccurs.size(); ++i) { - if (token_noccurs[i] == 0) continue; - ++n_unique_tokens; - } - printf("%s: number of unique tokens: %d\n", __func__, n_unique_tokens); - - size_t shuffle_samples_hash = compute_samples_hash(params.common.fn_train_data, train_samples_begin.data(), train_samples_size.data(), train_samples_size.size()); - const bool changed_train_data = (shuffle_samples_hash != train->shuffle_samples_hash) || (train->shuffle_sample_count != train_samples_size.size()); - if (changed_train_data) { - printf("%s: train data seems to have changed. restarting shuffled epoch.\n", __func__); - } - if (params.common.force_reshuffle) { - printf("%s: forced reshuffling of data. restarting with newly shuffled epoch.\n", __func__); - } - if ((train->shuffle_rng_state_current == "") || changed_train_data || params.common.force_reshuffle) { - train->shuffle_rng_state_current = mt19937_seed_to_state(params.common.seed); - train->shuffle_sample_count = train_samples_size.size(); - train->shuffle_next_sample = 0; - train->shuffle_samples_hash = shuffle_samples_hash; - } - std::vector train_shuffled_samples_offs; - std::vector train_shuffled_samples_begin; - std::vector train_shuffled_samples_size; - train_shuffled_samples_offs.resize(train_samples_begin.size()); - train_shuffled_samples_begin.resize(train_samples_begin.size()); - train_shuffled_samples_size.resize(train_samples_size.size()); - train->shuffle_rng_state_next = shuffle_samples( - train->shuffle_rng_state_current, - train_shuffled_samples_offs.data(), - train_shuffled_samples_begin.data(), - train_shuffled_samples_size.data(), - train_samples_begin.data(), - train_samples_size.data(), - train_samples_size.size()); - - printf("%s: begin training\n", __func__); - - save_train_files_data save_data; - save_data.fn_checkpoint_out = params.common.fn_checkpoint_out; - save_data.fn_lora_out = params.fn_lora_out; - save_data.pattern_fn_it = params.common.pattern_fn_it; - save_data.fn_latest = params.common.fn_latest; - save_data.model = &model; - save_data.lora = &lora; - - struct train_opt_callback_data opt_cb_data; - opt_cb_data.params = ¶ms.common; - opt_cb_data.train = train; - opt_cb_data.save_cb = &save_train_files; - opt_cb_data.save_data = &save_data; - opt_cb_data.lctx = lctx; - opt_cb_data.last_save_iter = opt->iter; - opt_cb_data.tokens_data = train_tokens.data(); - opt_cb_data.tokens_size = train_tokens.size(); - opt_cb_data.samples_begin = train_samples_begin.data(); - opt_cb_data.samples_size = train_samples_size.data(); - opt_cb_data.shuffled_samples_offs = train_shuffled_samples_offs.data(); - opt_cb_data.shuffled_samples_begin = train_shuffled_samples_begin.data(); - opt_cb_data.shuffled_samples_size = train_shuffled_samples_size.data(); - opt_cb_data.samples_count = train_samples_size.size(); - opt_cb_data.tokens_input = tokens_input; - opt_cb_data.target_probs = target_probs; - opt_cb_data.first_iter = opt->iter; - opt_cb_data.first_epoch = train->train_epochs; - opt_cb_data.iter_at_last_epoch = -1; - opt_cb_data.last_time = ggml_time_ms(); - opt_cb_data.millis_per_iter = 0.0; - - // measure required memory for work buffer - size_t max_work_size = ggml_graph_plan(gb, params.common.n_threads).work_size + GGML_OBJECT_SIZE; - printf("%s: work_size = %zu bytes (%.1f MB)\n", __func__, max_work_size, (float) max_work_size / (1024.0f*1024.0f)); - - // context for work buffer - struct ggml_init_params ctx_work_params = { - max_work_size, // mem_size - NULL, // mem_buffer - false, // no_alloc - }; - struct ggml_context * ctx_work = ggml_init(ctx_work_params); - - int64_t t0 = ggml_time_ms(); - - ggml_opt_resume_g(ctx_work, opt, loss, gf, gb, &train_opt_callback, (void *) &opt_cb_data); - - ggml_free(ctx_work); - ggml_free(ctx_compute); - ggml_free(ctx_input); - ggml_gallocr_free(alloc); - - - int64_t t1 = ggml_time_ms(); - printf("%s: total training time: ", __func__); - print_duration((double) (t1 - t0)); - printf("\n"); - - int new_iters = opt->iter - opt_cb_data.last_save_iter; - if (new_iters > 0) { - train->train_its += new_iters; - train->train_tokens += new_iters * opt->params.n_gradient_accumulation * n_batch * n_tokens; - - save_train_files(&save_data, train); - opt_cb_data.last_save_iter = opt->iter; - } - - ggml_free(opt->ctx); - free_train_state(train); - ggml_free(lora.ctx); - llama_free(lctx); - llama_free_model(lmodel); - return 0; -} diff --git a/examples/finetune/finetune.sh b/examples/finetune/finetune.sh deleted file mode 100644 index d7f2165e5..000000000 --- a/examples/finetune/finetune.sh +++ /dev/null @@ -1,34 +0,0 @@ -#!/bin/bash -cd `dirname $0` -cd ../.. - -EXE="./llama-finetune" - -if [[ ! $LLAMA_MODEL_DIR ]]; then LLAMA_MODEL_DIR="./models"; fi -if [[ ! $LLAMA_TRAINING_DIR ]]; then LLAMA_TRAINING_DIR="."; fi - -# MODEL="$LLAMA_MODEL_DIR/openllama-3b-v2-q8_0.gguf" # This is the model the readme uses. -MODEL="$LLAMA_MODEL_DIR/openllama-3b-v2.gguf" # An f16 model. Note in this case with "-g", you get an f32-format .BIN file that isn't yet supported if you use it with "main --lora" with GPU inferencing. - -while getopts "dg" opt; do - case $opt in - d) - DEBUGGER="gdb --args" - ;; - g) - EXE="./build/bin/Release/finetune" - GPUARG="--gpu-layers 25" - ;; - esac -done - -$DEBUGGER $EXE \ - --model-base $MODEL \ - $GPUARG \ - --checkpoint-in chk-ol3b-shakespeare-LATEST.gguf \ - --checkpoint-out chk-ol3b-shakespeare-ITERATION.gguf \ - --lora-out lora-ol3b-shakespeare-ITERATION.bin \ - --train-data "$LLAMA_TRAINING_DIR\shakespeare.txt" \ - --save-every 10 \ - --threads 10 --adam-iter 30 --batch 4 --ctx 64 \ - --use-checkpointing diff --git a/examples/gbnf-validator/gbnf-validator.cpp b/examples/gbnf-validator/gbnf-validator.cpp index dd53ba9b1..48a705e15 100644 --- a/examples/gbnf-validator/gbnf-validator.cpp +++ b/examples/gbnf-validator/gbnf-validator.cpp @@ -16,20 +16,25 @@ static bool llama_sample_grammar_string(struct llama_grammar * grammar, const st auto decoded = decode_utf8(input_str, {}); const auto & code_points = decoded.first; + const llama_grammar_rules & rules = llama_grammar_get_rules (grammar); + llama_grammar_stacks & cur_stacks = llama_grammar_get_stacks(grammar); + size_t pos = 0; for (auto it = code_points.begin(), end = code_points.end() - 1; it != end; ++it) { - auto prev_stacks = grammar->stacks; - llama_grammar_accept(grammar->rules, prev_stacks, *it, grammar->stacks); - if (grammar->stacks.empty()) { + const llama_grammar_stacks prev_stacks = llama_grammar_get_stacks(grammar); // copy + + llama_grammar_accept(rules, prev_stacks, *it, cur_stacks); + + if (cur_stacks.empty()) { error_pos = pos; error_msg = "Unexpected character '" + unicode_cpt_to_utf8(*it) + "'"; - grammar->stacks = prev_stacks; + cur_stacks = prev_stacks; return false; } ++pos; } - for (const auto & stack : grammar->stacks) { + for (const auto & stack : cur_stacks) { if (stack.empty()) { return true; } diff --git a/examples/gguf-hash/CMakeLists.txt b/examples/gguf-hash/CMakeLists.txt new file mode 100644 index 000000000..633f45535 --- /dev/null +++ b/examples/gguf-hash/CMakeLists.txt @@ -0,0 +1,15 @@ +set(TARGET llama-gguf-hash) +add_executable(${TARGET} gguf-hash.cpp) +install(TARGETS ${TARGET} RUNTIME) + +# clibs dependencies +include_directories(deps/) +add_library(xxhash OBJECT deps/xxhash/xxhash.c deps/xxhash/xxhash.h) +target_link_libraries(${TARGET} PRIVATE xxhash) +add_library(sha1 OBJECT deps/sha1/sha1.c deps/sha1/sha1.h) +target_link_libraries(${TARGET} PRIVATE sha1) +add_library(sha256 OBJECT deps/sha256/sha256.c deps/sha256/sha256.h) +target_link_libraries(${TARGET} PRIVATE sha256) + +target_link_libraries(${TARGET} PRIVATE ggml ${CMAKE_THREAD_LIBS_INIT}) +target_compile_features(${TARGET} PRIVATE cxx_std_11) diff --git a/examples/gguf-hash/README.md b/examples/gguf-hash/README.md new file mode 100644 index 000000000..9871651e3 --- /dev/null +++ b/examples/gguf-hash/README.md @@ -0,0 +1,206 @@ + +# llama-gguf-hash + +CLI to hash GGUF files to detect difference on a per model and per tensor level. + +**Command line options:** + +- `--help`: display help message +- `--xxh64`: use xhash 64bit hash mode (default) +- `--sha1`: use sha1 +- `--uuid`: use uuid +- `--sha256`: use sha256 +- `--all`: use all hash +- `--no-layer`: exclude per layer hash +- `--uuid`: generate UUIDv5 ID +- `-c`, `--check `: verify against a manifest + +## About + +While most POSIX systems already have hash checking programs like sha256sum, it +is designed to check entire files. This is not ideal for our purpose if we want +to check for consistency of the tensor data even if the metadata content of the +gguf KV store has been updated. + +This program is designed to hash a gguf tensor payload on a 'per tensor layer' +in addition to a 'entire tensor model' hash. The intent is that the entire +tensor layer can be checked first but if there is any detected inconsistencies, +then the per tensor hash can be used to narrow down the specific tensor layer +that has inconsistencies. + +For Maintainers: +- Detection of tensor inconsistency during development and automated tests + - This is served by xxh64 which is fast + - This is also served by having per tensor layer to assist in narrowing down + the location of the faulty tensor layer + - This is also served by sha1 which is much slower but more widely supported + +For Model Creators: +- Optional consistent UUID generation based on model tensor content + - This is served by UUIDv5 which is useful for databases keys + - llama.cpp UUIDv5 Namespace: `ef001206-dadc-5f6d-a15f-3359e577d4e5` + - Made via UUIDv5 URL namespace of `en.wikipedia.org/wiki/Llama.cpp` + +For Model Users: +- Assurance of tensor layer integrity even if metadata was updated + - This is served by sha256 which is still considered very secure as of 2024 + +### Design Note + +- The default behavior of this program if no arguments is provided is to hash + using xxhash's xxh32 mode because it is very fast and is primarily targeted + towards maintainers who may want to use this in automated tests. +- xxhash support xxh32 and xxh128 for 32bit hash and 128bit hash respectively + however we picked 64bit xxhash as most computers are 64bit as of 2024 and thus + would have a better affinity to calculating hash that is 64bit in size. + +## Compile Example + +```bash +cmake -B build -DCMAKE_BUILD_TYPE=Debug -DLLAMA_FATAL_WARNINGS=ON +make -C build clean +make -C build llama-gguf-hash VERBOSE=1 +./build/bin/llama-gguf-hash test.gguf +./build/bin/llama-gguf-hash --xxh64 test.gguf +./build/bin/llama-gguf-hash --sha1 test.gguf +./build/bin/llama-gguf-hash --uuid test.gguf +./build/bin/llama-gguf-hash --sha256 test.gguf +``` + +## Generation and Verification Example + +To generate we may use this command + +```bash +./llama-gguf-hash --all test.gguf > test.gguf.manifest +``` + +Which would generate a manifest that looks like below, which contains multiple hash type and per tensor layer hashes as well +(This excludes UUID as that is an ID not a hash) + +```bash +xxh64 f66e9cd66a4396a0 test.gguf:tensor_0 +sha1 59f79ecefd8125a996fdf419239051a7e99e5f20 test.gguf:tensor_0 +sha256 c0510d38fa060c46265e0160a85c7243096b01dd31c2f355bdbb5516b20de1bd test.gguf:tensor_0 +xxh64 7d3a1f9ac04d0537 test.gguf:tensor_1 +sha1 4765f592eacf096df4628ba59476af94d767080a test.gguf:tensor_1 +sha256 8514cbcc73692a2c56bd7a33a022edd5ff819614bd23b19915d7224387f397a7 test.gguf:tensor_1 +xxh64 a0af5d700049693b test.gguf:tensor_2 +sha1 25cbfbad4513cc348e2c95ebdee69d6ff2fd8753 test.gguf:tensor_2 +sha256 947e6b36e20f2cc95e1d2ce1c1669d813d574657ac6b5ac5196158d454d35180 test.gguf:tensor_2 +xxh64 e83fddf559d7b6a6 test.gguf:tensor_3 +sha1 a9cba73e2d90f2ee3dae2548caa42bef3fe6a96c test.gguf:tensor_3 +sha256 423b044e016d8ac73c39f23f60bf01bedef5ecb03c0230accd824c91fe86f1a1 test.gguf:tensor_3 +xxh64 1257733306b7992d test.gguf:tensor_4 +sha1 d7bc61db93bb685ce9d598da89717c66729b7543 test.gguf:tensor_4 +sha256 79737cb3912d4201384cf7f16a1a37ff7823f23ea796cb205b6ca361ab9e3ebf test.gguf:tensor_4 +xxh64 d238d16ba4711e58 test.gguf:tensor_5 +sha1 0706566c198fe1072f37e0a5135b4b5f23654c52 test.gguf:tensor_5 +sha256 60949be8298eced0ecdde64487643d018407bd261691e061d9e9c3dbc9fd358b test.gguf:tensor_5 +xxh64 3fbc3b65ab8c7f39 test.gguf:tensor_6 +sha1 73922a0727226a409049f6fc3172a52219ca6f00 test.gguf:tensor_6 +sha256 574f4c46ff384a3b9a225eb955d2a871847a2e8b3fa59387a8252832e92ef7b0 test.gguf:tensor_6 +xxh64 c22021c29854f093 test.gguf:tensor_7 +sha1 efc39cece6a951188fc41e354c73bbfe6813d447 test.gguf:tensor_7 +sha256 4c0410cd3c500f078ae5b21e8dc9eb79e29112713b2ab58a882f82a3868d4d75 test.gguf:tensor_7 +xxh64 936df61f5d64261f test.gguf:tensor_8 +sha1 c2490296d789a4f34398a337fed8377d943d9f06 test.gguf:tensor_8 +sha256 c4401313feeba0261275c3b25bd2d8fe40ce04e0f440c2980ed0e9674c30ff01 test.gguf:tensor_8 +xxh64 93fd20c64421c081 test.gguf:tensor_9 +sha1 7047ce1e78437a6884337a3751c7ee0421918a65 test.gguf:tensor_9 +sha256 23d57cf0d7a6e90b0b3616b41300e0cd354781e812add854a5f95aa55f2bc514 test.gguf:tensor_9 +xxh64 5a54d3aad816f302 test.gguf +sha1 d15be52c4ff213e823cb6dd13af7ee2f978e7042 test.gguf +sha256 7dd641b32f59b60dbd4b5420c4b0f6321ccf48f58f6ae201a3dbc4a58a27c6e4 test.gguf +``` + +We can then use the normal check command which will by default check for the highest security strength hash and verify against that: + +```bash +$ ./llama-gguf-hash --check test.gguf.manifest test.gguf +manifest test.gguf.manifest sha256 sha1 xxh64 +sha256 c0510d38fa060c46265e0160a85c7243096b01dd31c2f355bdbb5516b20de1bd test.gguf:tensor_0 - Ok +sha256 8514cbcc73692a2c56bd7a33a022edd5ff819614bd23b19915d7224387f397a7 test.gguf:tensor_1 - Ok +sha256 947e6b36e20f2cc95e1d2ce1c1669d813d574657ac6b5ac5196158d454d35180 test.gguf:tensor_2 - Ok +sha256 423b044e016d8ac73c39f23f60bf01bedef5ecb03c0230accd824c91fe86f1a1 test.gguf:tensor_3 - Ok +sha256 79737cb3912d4201384cf7f16a1a37ff7823f23ea796cb205b6ca361ab9e3ebf test.gguf:tensor_4 - Ok +sha256 60949be8298eced0ecdde64487643d018407bd261691e061d9e9c3dbc9fd358b test.gguf:tensor_5 - Ok +sha256 574f4c46ff384a3b9a225eb955d2a871847a2e8b3fa59387a8252832e92ef7b0 test.gguf:tensor_6 - Ok +sha256 4c0410cd3c500f078ae5b21e8dc9eb79e29112713b2ab58a882f82a3868d4d75 test.gguf:tensor_7 - Ok +sha256 c4401313feeba0261275c3b25bd2d8fe40ce04e0f440c2980ed0e9674c30ff01 test.gguf:tensor_8 - Ok +sha256 23d57cf0d7a6e90b0b3616b41300e0cd354781e812add854a5f95aa55f2bc514 test.gguf:tensor_9 - Ok +sha256 7dd641b32f59b60dbd4b5420c4b0f6321ccf48f58f6ae201a3dbc4a58a27c6e4 test.gguf - Ok + +Verification results for test.gguf.manifest - Success +``` + +Or we may explicitly ask for a faster hash like: + +```bash +$ ./llama-gguf-hash --check test.gguf.manifest --xxh64 test.gguf +manifest test.gguf.manifest sha256 sha1 xxh64 +xxh64 f66e9cd66a4396a0 test.gguf:tensor_0 - Ok +xxh64 7d3a1f9ac04d0537 test.gguf:tensor_1 - Ok +xxh64 a0af5d700049693b test.gguf:tensor_2 - Ok +xxh64 e83fddf559d7b6a6 test.gguf:tensor_3 - Ok +xxh64 1257733306b7992d test.gguf:tensor_4 - Ok +xxh64 d238d16ba4711e58 test.gguf:tensor_5 - Ok +xxh64 3fbc3b65ab8c7f39 test.gguf:tensor_6 - Ok +xxh64 c22021c29854f093 test.gguf:tensor_7 - Ok +xxh64 936df61f5d64261f test.gguf:tensor_8 - Ok +xxh64 93fd20c64421c081 test.gguf:tensor_9 - Ok +xxh64 5a54d3aad816f302 test.gguf - Ok + +Verification results for test.gguf.manifest - Success +``` + +Or maybe we want to just check that all the hash is valid: + +```bash +$./llama-gguf-hash --check test.gguf.manifest --all test.gguf.manifest +manifest test.gguf.manifest sha256 sha1 xxh64 +xxh64 f66e9cd66a4396a0 test.gguf:tensor_0 - Ok +sha1 59f79ecefd8125a996fdf419239051a7e99e5f20 test.gguf:tensor_0 - Ok +sha256 c0510d38fa060c46265e0160a85c7243096b01dd31c2f355bdbb5516b20de1bd test.gguf:tensor_0 - Ok +xxh64 7d3a1f9ac04d0537 test.gguf:tensor_1 - Ok +sha1 4765f592eacf096df4628ba59476af94d767080a test.gguf:tensor_1 - Ok +sha256 8514cbcc73692a2c56bd7a33a022edd5ff819614bd23b19915d7224387f397a7 test.gguf:tensor_1 - Ok +xxh64 a0af5d700049693b test.gguf:tensor_2 - Ok +sha1 25cbfbad4513cc348e2c95ebdee69d6ff2fd8753 test.gguf:tensor_2 - Ok +sha256 947e6b36e20f2cc95e1d2ce1c1669d813d574657ac6b5ac5196158d454d35180 test.gguf:tensor_2 - Ok +xxh64 e83fddf559d7b6a6 test.gguf:tensor_3 - Ok +sha1 a9cba73e2d90f2ee3dae2548caa42bef3fe6a96c test.gguf:tensor_3 - Ok +sha256 423b044e016d8ac73c39f23f60bf01bedef5ecb03c0230accd824c91fe86f1a1 test.gguf:tensor_3 - Ok +xxh64 1257733306b7992d test.gguf:tensor_4 - Ok +sha1 d7bc61db93bb685ce9d598da89717c66729b7543 test.gguf:tensor_4 - Ok +sha256 79737cb3912d4201384cf7f16a1a37ff7823f23ea796cb205b6ca361ab9e3ebf test.gguf:tensor_4 - Ok +xxh64 d238d16ba4711e58 test.gguf:tensor_5 - Ok +sha1 0706566c198fe1072f37e0a5135b4b5f23654c52 test.gguf:tensor_5 - Ok +sha256 60949be8298eced0ecdde64487643d018407bd261691e061d9e9c3dbc9fd358b test.gguf:tensor_5 - Ok +xxh64 3fbc3b65ab8c7f39 test.gguf:tensor_6 - Ok +sha1 73922a0727226a409049f6fc3172a52219ca6f00 test.gguf:tensor_6 - Ok +sha256 574f4c46ff384a3b9a225eb955d2a871847a2e8b3fa59387a8252832e92ef7b0 test.gguf:tensor_6 - Ok +xxh64 c22021c29854f093 test.gguf:tensor_7 - Ok +sha1 efc39cece6a951188fc41e354c73bbfe6813d447 test.gguf:tensor_7 - Ok +sha256 4c0410cd3c500f078ae5b21e8dc9eb79e29112713b2ab58a882f82a3868d4d75 test.gguf:tensor_7 - Ok +xxh64 936df61f5d64261f test.gguf:tensor_8 - Ok +sha1 c2490296d789a4f34398a337fed8377d943d9f06 test.gguf:tensor_8 - Ok +sha256 c4401313feeba0261275c3b25bd2d8fe40ce04e0f440c2980ed0e9674c30ff01 test.gguf:tensor_8 - Ok +xxh64 93fd20c64421c081 test.gguf:tensor_9 - Ok +sha1 7047ce1e78437a6884337a3751c7ee0421918a65 test.gguf:tensor_9 - Ok +sha256 23d57cf0d7a6e90b0b3616b41300e0cd354781e812add854a5f95aa55f2bc514 test.gguf:tensor_9 - Ok +xxh64 5a54d3aad816f302 test.gguf - Ok +sha1 d15be52c4ff213e823cb6dd13af7ee2f978e7042 test.gguf - Ok +sha256 7dd641b32f59b60dbd4b5420c4b0f6321ccf48f58f6ae201a3dbc4a58a27c6e4 test.gguf - Ok + +Verification results for test.gguf.manifest - Success +``` + + +## Crypto/Hash Libraries Used + +These micro c libraries dependencies was installed via the [clib c package manager](https://github.com/clibs) + +- https://github.com/Cyan4973/xxHash +- https://github.com/clibs/sha1/ +- https://github.com/jb55/sha256.c diff --git a/examples/gguf-hash/deps/rotate-bits/package.json b/examples/gguf-hash/deps/rotate-bits/package.json new file mode 100644 index 000000000..74c0bef68 --- /dev/null +++ b/examples/gguf-hash/deps/rotate-bits/package.json @@ -0,0 +1,13 @@ +{ + "name": "rotate-bits", + "version": "0.1.1", + "repo": "jb55/rotate-bits.h", + "description": "rotate bits", + "keywords": ["rotl", "rotr"], + "src": ["rotate-bits.h"], + "license": "Public Domain", + "development": { + "thlorenz/tap.c": "*" + } +} + diff --git a/examples/gguf-hash/deps/rotate-bits/rotate-bits.h b/examples/gguf-hash/deps/rotate-bits/rotate-bits.h new file mode 100644 index 000000000..75c4881fc --- /dev/null +++ b/examples/gguf-hash/deps/rotate-bits/rotate-bits.h @@ -0,0 +1,46 @@ + + +#ifndef __ROTATE_DEFS_H +#define __ROTATE_DEFS_H + +#ifdef _MSC_VER + +#include + +#define ROTL32(v, n) _rotl((v), (n)) +#define ROTL64(v, n) _rotl64((v), (n)) + +#define ROTR32(v, n) _rotr((v), (n)) +#define ROTR64(v, n) _rotr64((v), (n)) + +#else + +#include + +#define U8V(v) ((uint8_t)(v) & 0xFFU) +#define U16V(v) ((uint16_t)(v) & 0xFFFFU) +#define U32V(v) ((uint32_t)(v) & 0xFFFFFFFFU) +#define U64V(v) ((uint64_t)(v) & 0xFFFFFFFFFFFFFFFFU) + +#define ROTL32(v, n) \ + (U32V((uint32_t)(v) << (n)) | ((uint32_t)(v) >> (32 - (n)))) + +// tests fail if we don't have this cast... +#define ROTL64(v, n) \ + (U64V((uint64_t)(v) << (n)) | ((uint64_t)(v) >> (64 - (n)))) + +#define ROTR32(v, n) ROTL32(v, 32 - (n)) +#define ROTR64(v, n) ROTL64(v, 64 - (n)) + +#endif + +#define ROTL8(v, n) \ + (U8V((uint8_t)(v) << (n)) | ((uint8_t)(v) >> (8 - (n)))) + +#define ROTL16(v, n) \ + (U16V((uint16_t)(v) << (n)) | ((uint16_t)(v) >> (16 - (n)))) + +#define ROTR8(v, n) ROTL8(v, 8 - (n)) +#define ROTR16(v, n) ROTL16(v, 16 - (n)) + +#endif diff --git a/examples/gguf-hash/deps/sha1/package.json b/examples/gguf-hash/deps/sha1/package.json new file mode 100644 index 000000000..6a5843dd1 --- /dev/null +++ b/examples/gguf-hash/deps/sha1/package.json @@ -0,0 +1,9 @@ +{ + "name": "sha1", + "version": "0.0.1", + "repo": "clibs/sha1", + "description": "sha1 hash algorithm", + "keywords": ["sha1", "hash"], + "license": "public domain", + "src": ["sha1.c", "sha1.h"] +} diff --git a/examples/gguf-hash/deps/sha1/sha1.c b/examples/gguf-hash/deps/sha1/sha1.c new file mode 100644 index 000000000..76cd6ca33 --- /dev/null +++ b/examples/gguf-hash/deps/sha1/sha1.c @@ -0,0 +1,295 @@ +/* +SHA-1 in C +By Steve Reid +100% Public Domain + +Test Vectors (from FIPS PUB 180-1) +"abc" + A9993E36 4706816A BA3E2571 7850C26C 9CD0D89D +"abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq" + 84983E44 1C3BD26E BAAE4AA1 F95129E5 E54670F1 +A million repetitions of "a" + 34AA973C D4C4DAA4 F61EEB2B DBAD2731 6534016F +*/ + +/* #define LITTLE_ENDIAN * This should be #define'd already, if true. */ +/* #define SHA1HANDSOFF * Copies data before messing with it. */ + +#define SHA1HANDSOFF + +#include +#include + +/* for uint32_t */ +#include + +#include "sha1.h" + + +#define rol(value, bits) (((value) << (bits)) | ((value) >> (32 - (bits)))) + +/* blk0() and blk() perform the initial expand. */ +/* I got the idea of expanding during the round function from SSLeay */ +#if BYTE_ORDER == LITTLE_ENDIAN +#define blk0(i) (block->l[i] = (rol(block->l[i],24)&0xFF00FF00) \ + |(rol(block->l[i],8)&0x00FF00FF)) +#elif BYTE_ORDER == BIG_ENDIAN +#define blk0(i) block->l[i] +#else +#error "Endianness not defined!" +#endif +#define blk(i) (block->l[i&15] = rol(block->l[(i+13)&15]^block->l[(i+8)&15] \ + ^block->l[(i+2)&15]^block->l[i&15],1)) + +/* (R0+R1), R2, R3, R4 are the different operations used in SHA1 */ +#define R0(v,w,x,y,z,i) z+=((w&(x^y))^y)+blk0(i)+0x5A827999+rol(v,5);w=rol(w,30); +#define R1(v,w,x,y,z,i) z+=((w&(x^y))^y)+blk(i)+0x5A827999+rol(v,5);w=rol(w,30); +#define R2(v,w,x,y,z,i) z+=(w^x^y)+blk(i)+0x6ED9EBA1+rol(v,5);w=rol(w,30); +#define R3(v,w,x,y,z,i) z+=(((w|x)&y)|(w&x))+blk(i)+0x8F1BBCDC+rol(v,5);w=rol(w,30); +#define R4(v,w,x,y,z,i) z+=(w^x^y)+blk(i)+0xCA62C1D6+rol(v,5);w=rol(w,30); + + +/* Hash a single 512-bit block. This is the core of the algorithm. */ + +void SHA1Transform( + uint32_t state[5], + const unsigned char buffer[64] +) +{ + uint32_t a, b, c, d, e; + + typedef union + { + unsigned char c[64]; + uint32_t l[16]; + } CHAR64LONG16; + +#ifdef SHA1HANDSOFF + CHAR64LONG16 block[1]; /* use array to appear as a pointer */ + + memcpy(block, buffer, 64); +#else + /* The following had better never be used because it causes the + * pointer-to-const buffer to be cast into a pointer to non-const. + * And the result is written through. I threw a "const" in, hoping + * this will cause a diagnostic. + */ + CHAR64LONG16 *block = (const CHAR64LONG16 *) buffer; +#endif + /* Copy context->state[] to working vars */ + a = state[0]; + b = state[1]; + c = state[2]; + d = state[3]; + e = state[4]; + /* 4 rounds of 20 operations each. Loop unrolled. */ + R0(a, b, c, d, e, 0); + R0(e, a, b, c, d, 1); + R0(d, e, a, b, c, 2); + R0(c, d, e, a, b, 3); + R0(b, c, d, e, a, 4); + R0(a, b, c, d, e, 5); + R0(e, a, b, c, d, 6); + R0(d, e, a, b, c, 7); + R0(c, d, e, a, b, 8); + R0(b, c, d, e, a, 9); + R0(a, b, c, d, e, 10); + R0(e, a, b, c, d, 11); + R0(d, e, a, b, c, 12); + R0(c, d, e, a, b, 13); + R0(b, c, d, e, a, 14); + R0(a, b, c, d, e, 15); + R1(e, a, b, c, d, 16); + R1(d, e, a, b, c, 17); + R1(c, d, e, a, b, 18); + R1(b, c, d, e, a, 19); + R2(a, b, c, d, e, 20); + R2(e, a, b, c, d, 21); + R2(d, e, a, b, c, 22); + R2(c, d, e, a, b, 23); + R2(b, c, d, e, a, 24); + R2(a, b, c, d, e, 25); + R2(e, a, b, c, d, 26); + R2(d, e, a, b, c, 27); + R2(c, d, e, a, b, 28); + R2(b, c, d, e, a, 29); + R2(a, b, c, d, e, 30); + R2(e, a, b, c, d, 31); + R2(d, e, a, b, c, 32); + R2(c, d, e, a, b, 33); + R2(b, c, d, e, a, 34); + R2(a, b, c, d, e, 35); + R2(e, a, b, c, d, 36); + R2(d, e, a, b, c, 37); + R2(c, d, e, a, b, 38); + R2(b, c, d, e, a, 39); + R3(a, b, c, d, e, 40); + R3(e, a, b, c, d, 41); + R3(d, e, a, b, c, 42); + R3(c, d, e, a, b, 43); + R3(b, c, d, e, a, 44); + R3(a, b, c, d, e, 45); + R3(e, a, b, c, d, 46); + R3(d, e, a, b, c, 47); + R3(c, d, e, a, b, 48); + R3(b, c, d, e, a, 49); + R3(a, b, c, d, e, 50); + R3(e, a, b, c, d, 51); + R3(d, e, a, b, c, 52); + R3(c, d, e, a, b, 53); + R3(b, c, d, e, a, 54); + R3(a, b, c, d, e, 55); + R3(e, a, b, c, d, 56); + R3(d, e, a, b, c, 57); + R3(c, d, e, a, b, 58); + R3(b, c, d, e, a, 59); + R4(a, b, c, d, e, 60); + R4(e, a, b, c, d, 61); + R4(d, e, a, b, c, 62); + R4(c, d, e, a, b, 63); + R4(b, c, d, e, a, 64); + R4(a, b, c, d, e, 65); + R4(e, a, b, c, d, 66); + R4(d, e, a, b, c, 67); + R4(c, d, e, a, b, 68); + R4(b, c, d, e, a, 69); + R4(a, b, c, d, e, 70); + R4(e, a, b, c, d, 71); + R4(d, e, a, b, c, 72); + R4(c, d, e, a, b, 73); + R4(b, c, d, e, a, 74); + R4(a, b, c, d, e, 75); + R4(e, a, b, c, d, 76); + R4(d, e, a, b, c, 77); + R4(c, d, e, a, b, 78); + R4(b, c, d, e, a, 79); + /* Add the working vars back into context.state[] */ + state[0] += a; + state[1] += b; + state[2] += c; + state[3] += d; + state[4] += e; + /* Wipe variables */ + a = b = c = d = e = 0; +#ifdef SHA1HANDSOFF + memset(block, '\0', sizeof(block)); +#endif +} + + +/* SHA1Init - Initialize new context */ + +void SHA1Init( + SHA1_CTX * context +) +{ + /* SHA1 initialization constants */ + context->state[0] = 0x67452301; + context->state[1] = 0xEFCDAB89; + context->state[2] = 0x98BADCFE; + context->state[3] = 0x10325476; + context->state[4] = 0xC3D2E1F0; + context->count[0] = context->count[1] = 0; +} + + +/* Run your data through this. */ + +void SHA1Update( + SHA1_CTX * context, + const unsigned char *data, + uint32_t len +) +{ + uint32_t i; + + uint32_t j; + + j = context->count[0]; + if ((context->count[0] += len << 3) < j) + context->count[1]++; + context->count[1] += (len >> 29); + j = (j >> 3) & 63; + if ((j + len) > 63) + { + memcpy(&context->buffer[j], data, (i = 64 - j)); + SHA1Transform(context->state, context->buffer); + for (; i + 63 < len; i += 64) + { + SHA1Transform(context->state, &data[i]); + } + j = 0; + } + else + i = 0; + memcpy(&context->buffer[j], &data[i], len - i); +} + + +/* Add padding and return the message digest. */ + +void SHA1Final( + unsigned char digest[20], + SHA1_CTX * context +) +{ + unsigned i; + + unsigned char finalcount[8]; + + unsigned char c; + +#if 0 /* untested "improvement" by DHR */ + /* Convert context->count to a sequence of bytes + * in finalcount. Second element first, but + * big-endian order within element. + * But we do it all backwards. + */ + unsigned char *fcp = &finalcount[8]; + + for (i = 0; i < 2; i++) + { + uint32_t t = context->count[i]; + + int j; + + for (j = 0; j < 4; t >>= 8, j++) + *--fcp = (unsigned char) t} +#else + for (i = 0; i < 8; i++) + { + finalcount[i] = (unsigned char) ((context->count[(i >= 4 ? 0 : 1)] >> ((3 - (i & 3)) * 8)) & 255); /* Endian independent */ + } +#endif + c = 0200; + SHA1Update(context, &c, 1); + while ((context->count[0] & 504) != 448) + { + c = 0000; + SHA1Update(context, &c, 1); + } + SHA1Update(context, finalcount, 8); /* Should cause a SHA1Transform() */ + for (i = 0; i < 20; i++) + { + digest[i] = (unsigned char) + ((context->state[i >> 2] >> ((3 - (i & 3)) * 8)) & 255); + } + /* Wipe variables */ + memset(context, '\0', sizeof(*context)); + memset(&finalcount, '\0', sizeof(finalcount)); +} + +void SHA1( + char *hash_out, + const char *str, + uint32_t len) +{ + SHA1_CTX ctx; + unsigned int ii; + + SHA1Init(&ctx); + for (ii=0; ii + 100% Public Domain + */ + +#include "stdint.h" + +#if defined(__cplusplus) +extern "C" { +#endif + +typedef struct +{ + uint32_t state[5]; + uint32_t count[2]; + unsigned char buffer[64]; +} SHA1_CTX; + +void SHA1Transform( + uint32_t state[5], + const unsigned char buffer[64] + ); + +void SHA1Init( + SHA1_CTX * context + ); + +void SHA1Update( + SHA1_CTX * context, + const unsigned char *data, + uint32_t len + ); + +void SHA1Final( + unsigned char digest[20], + SHA1_CTX * context + ); + +void SHA1( + char *hash_out, + const char *str, + uint32_t len); + +#if defined(__cplusplus) +} +#endif + +#endif /* SHA1_H */ diff --git a/examples/gguf-hash/deps/sha256/package.json b/examples/gguf-hash/deps/sha256/package.json new file mode 100644 index 000000000..b92a04127 --- /dev/null +++ b/examples/gguf-hash/deps/sha256/package.json @@ -0,0 +1,15 @@ +{ + "name": "sha256", + "version": "0.0.2", + "repo": "jb55/sha256.c", + "description": "sha256 in c", + "keywords": ["sha256", "sha2"], + "src": ["sha256.c", "sha256.h"], + "dependencies": { + "jb55/rotate-bits.h": "0.1.1" + }, + "development": { + "thlorenz/tap.c": "*" + } +} + diff --git a/examples/gguf-hash/deps/sha256/sha256.c b/examples/gguf-hash/deps/sha256/sha256.c new file mode 100644 index 000000000..a7a87aeb2 --- /dev/null +++ b/examples/gguf-hash/deps/sha256/sha256.c @@ -0,0 +1,221 @@ +/* Crypto/Sha256.c -- SHA-256 Hash +2010-06-11 : Igor Pavlov : Public domain +This code is based on public domain code from Wei Dai's Crypto++ library. */ + +#include "rotate-bits/rotate-bits.h" +#include "sha256.h" + +/* define it for speed optimization */ +#define _SHA256_UNROLL +#define _SHA256_UNROLL2 + +void +sha256_init(sha256_t *p) +{ + p->state[0] = 0x6a09e667; + p->state[1] = 0xbb67ae85; + p->state[2] = 0x3c6ef372; + p->state[3] = 0xa54ff53a; + p->state[4] = 0x510e527f; + p->state[5] = 0x9b05688c; + p->state[6] = 0x1f83d9ab; + p->state[7] = 0x5be0cd19; + p->count = 0; +} + +#define S0(x) (ROTR32(x, 2) ^ ROTR32(x,13) ^ ROTR32(x, 22)) +#define S1(x) (ROTR32(x, 6) ^ ROTR32(x,11) ^ ROTR32(x, 25)) +#define s0(x) (ROTR32(x, 7) ^ ROTR32(x,18) ^ (x >> 3)) +#define s1(x) (ROTR32(x,17) ^ ROTR32(x,19) ^ (x >> 10)) + +#define blk0(i) (W[i] = data[i]) +#define blk2(i) (W[i&15] += s1(W[(i-2)&15]) + W[(i-7)&15] + s0(W[(i-15)&15])) + +#define Ch(x,y,z) (z^(x&(y^z))) +#define Maj(x,y,z) ((x&y)|(z&(x|y))) + +#define a(i) T[(0-(i))&7] +#define b(i) T[(1-(i))&7] +#define c(i) T[(2-(i))&7] +#define d(i) T[(3-(i))&7] +#define e(i) T[(4-(i))&7] +#define f(i) T[(5-(i))&7] +#define g(i) T[(6-(i))&7] +#define h(i) T[(7-(i))&7] + + +#ifdef _SHA256_UNROLL2 + +#define R(a,b,c,d,e,f,g,h, i) h += S1(e) + Ch(e,f,g) + K[i+j] + (j?blk2(i):blk0(i));\ + d += h; h += S0(a) + Maj(a, b, c) + +#define RX_8(i) \ + R(a,b,c,d,e,f,g,h, i); \ + R(h,a,b,c,d,e,f,g, (i+1)); \ + R(g,h,a,b,c,d,e,f, (i+2)); \ + R(f,g,h,a,b,c,d,e, (i+3)); \ + R(e,f,g,h,a,b,c,d, (i+4)); \ + R(d,e,f,g,h,a,b,c, (i+5)); \ + R(c,d,e,f,g,h,a,b, (i+6)); \ + R(b,c,d,e,f,g,h,a, (i+7)) + +#else + +#define R(i) h(i) += S1(e(i)) + Ch(e(i),f(i),g(i)) + K[i+j] + (j?blk2(i):blk0(i));\ + d(i) += h(i); h(i) += S0(a(i)) + Maj(a(i), b(i), c(i)) + +#ifdef _SHA256_UNROLL + +#define RX_8(i) R(i+0); R(i+1); R(i+2); R(i+3); R(i+4); R(i+5); R(i+6); R(i+7); + +#endif + +#endif + +static const uint32_t K[64] = { + 0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5, + 0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5, + 0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3, + 0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174, + 0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc, + 0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da, + 0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7, + 0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967, + 0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13, + 0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85, + 0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3, + 0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070, + 0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5, + 0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3, + 0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208, + 0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2 +}; + +static void +sha256_transform(uint32_t *state, const uint32_t *data) +{ + uint32_t W[16] = {0}; + unsigned j; + #ifdef _SHA256_UNROLL2 + uint32_t a,b,c,d,e,f,g,h; + a = state[0]; + b = state[1]; + c = state[2]; + d = state[3]; + e = state[4]; + f = state[5]; + g = state[6]; + h = state[7]; + #else + uint32_t T[8]; + for (j = 0; j < 8; j++) + T[j] = state[j]; + #endif + + for (j = 0; j < 64; j += 16) + { + #if defined(_SHA256_UNROLL) || defined(_SHA256_UNROLL2) + RX_8(0); RX_8(8); + #else + unsigned i; + for (i = 0; i < 16; i++) { R(i); } + #endif + } + + #ifdef _SHA256_UNROLL2 + state[0] += a; + state[1] += b; + state[2] += c; + state[3] += d; + state[4] += e; + state[5] += f; + state[6] += g; + state[7] += h; + #else + for (j = 0; j < 8; j++) + state[j] += T[j]; + #endif + + /* Wipe variables */ + /* memset(W, 0, sizeof(W)); */ + /* memset(T, 0, sizeof(T)); */ +} + +#undef S0 +#undef S1 +#undef s0 +#undef s1 + +static void +sha256_write_byte_block(sha256_t *p) +{ + uint32_t data32[16]; + unsigned i; + for (i = 0; i < 16; i++) + data32[i] = + ((uint32_t)(p->buffer[i * 4 ]) << 24) + + ((uint32_t)(p->buffer[i * 4 + 1]) << 16) + + ((uint32_t)(p->buffer[i * 4 + 2]) << 8) + + ((uint32_t)(p->buffer[i * 4 + 3])); + sha256_transform(p->state, data32); +} + + +void +sha256_hash(unsigned char *buf, const unsigned char *data, size_t size) +{ + sha256_t hash; + sha256_init(&hash); + sha256_update(&hash, data, size); + sha256_final(&hash, buf); +} + + +void +sha256_update(sha256_t *p, const unsigned char *data, size_t size) +{ + uint32_t curBufferPos = (uint32_t)p->count & 0x3F; + while (size > 0) + { + p->buffer[curBufferPos++] = *data++; + p->count++; + size--; + if (curBufferPos == 64) + { + curBufferPos = 0; + sha256_write_byte_block(p); + } + } +} + + +void +sha256_final(sha256_t *p, unsigned char *digest) +{ + uint64_t lenInBits = (p->count << 3); + uint32_t curBufferPos = (uint32_t)p->count & 0x3F; + unsigned i; + p->buffer[curBufferPos++] = 0x80; + while (curBufferPos != (64 - 8)) + { + curBufferPos &= 0x3F; + if (curBufferPos == 0) + sha256_write_byte_block(p); + p->buffer[curBufferPos++] = 0; + } + for (i = 0; i < 8; i++) + { + p->buffer[curBufferPos++] = (unsigned char)(lenInBits >> 56); + lenInBits <<= 8; + } + sha256_write_byte_block(p); + + for (i = 0; i < 8; i++) + { + *digest++ = (unsigned char)(p->state[i] >> 24); + *digest++ = (unsigned char)(p->state[i] >> 16); + *digest++ = (unsigned char)(p->state[i] >> 8); + *digest++ = (unsigned char)(p->state[i]); + } + sha256_init(p); +} diff --git a/examples/gguf-hash/deps/sha256/sha256.h b/examples/gguf-hash/deps/sha256/sha256.h new file mode 100644 index 000000000..21657e66b --- /dev/null +++ b/examples/gguf-hash/deps/sha256/sha256.h @@ -0,0 +1,24 @@ +/* Sha256.h -- SHA-256 Hash +2010-06-11 : Igor Pavlov : Public domain */ + +#ifndef __CRYPTO_SHA256_H +#define __CRYPTO_SHA256_H + +#include +#include + +#define SHA256_DIGEST_SIZE 32 + +typedef struct sha256_t +{ + uint32_t state[8]; + uint64_t count; + unsigned char buffer[64]; +} sha256_t; + +void sha256_init(sha256_t *p); +void sha256_update(sha256_t *p, const unsigned char *data, size_t size); +void sha256_final(sha256_t *p, unsigned char *digest); +void sha256_hash(unsigned char *buf, const unsigned char *data, size_t size); + +#endif diff --git a/examples/gguf-hash/deps/xxhash/clib.json b/examples/gguf-hash/deps/xxhash/clib.json new file mode 100644 index 000000000..242343c5d --- /dev/null +++ b/examples/gguf-hash/deps/xxhash/clib.json @@ -0,0 +1,12 @@ +{ + "name": "xxhash", + "version": "0.8.2", + "repo": "Cyan4973/xxhash", + "description": "Extremely fast non-cryptographic hash algorithm", + "keywords": ["xxhash", "hashing"], + "license": "BSD-2-Clause", + "src": [ + "xxhash.c", + "xxhash.h" + ] +} diff --git a/examples/gguf-hash/deps/xxhash/xxhash.c b/examples/gguf-hash/deps/xxhash/xxhash.c new file mode 100644 index 000000000..e60cc37f1 --- /dev/null +++ b/examples/gguf-hash/deps/xxhash/xxhash.c @@ -0,0 +1,42 @@ +/* + * xxHash - Extremely Fast Hash algorithm + * Copyright (C) 2012-2023 Yann Collet + * + * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are + * met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following disclaimer + * in the documentation and/or other materials provided with the + * distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + * + * You can contact the author at: + * - xxHash homepage: https://www.xxhash.com + * - xxHash source repository: https://github.com/Cyan4973/xxHash + */ + +/* + * xxhash.c instantiates functions defined in xxhash.h + */ + +#define XXH_STATIC_LINKING_ONLY /* access advanced declarations */ +#define XXH_IMPLEMENTATION /* access definitions */ + +#include "xxhash.h" diff --git a/examples/gguf-hash/deps/xxhash/xxhash.h b/examples/gguf-hash/deps/xxhash/xxhash.h new file mode 100644 index 000000000..c0fafe20d --- /dev/null +++ b/examples/gguf-hash/deps/xxhash/xxhash.h @@ -0,0 +1,7093 @@ +/* + * xxHash - Extremely Fast Hash algorithm + * Header File + * Copyright (C) 2012-2023 Yann Collet + * + * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are + * met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following disclaimer + * in the documentation and/or other materials provided with the + * distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + * + * You can contact the author at: + * - xxHash homepage: https://www.xxhash.com + * - xxHash source repository: https://github.com/Cyan4973/xxHash + */ + +/*! + * @mainpage xxHash + * + * xxHash is an extremely fast non-cryptographic hash algorithm, working at RAM speed + * limits. + * + * It is proposed in four flavors, in three families: + * 1. @ref XXH32_family + * - Classic 32-bit hash function. Simple, compact, and runs on almost all + * 32-bit and 64-bit systems. + * 2. @ref XXH64_family + * - Classic 64-bit adaptation of XXH32. Just as simple, and runs well on most + * 64-bit systems (but _not_ 32-bit systems). + * 3. @ref XXH3_family + * - Modern 64-bit and 128-bit hash function family which features improved + * strength and performance across the board, especially on smaller data. + * It benefits greatly from SIMD and 64-bit without requiring it. + * + * Benchmarks + * --- + * The reference system uses an Intel i7-9700K CPU, and runs Ubuntu x64 20.04. + * The open source benchmark program is compiled with clang v10.0 using -O3 flag. + * + * | Hash Name | ISA ext | Width | Large Data Speed | Small Data Velocity | + * | -------------------- | ------- | ----: | ---------------: | ------------------: | + * | XXH3_64bits() | @b AVX2 | 64 | 59.4 GB/s | 133.1 | + * | MeowHash | AES-NI | 128 | 58.2 GB/s | 52.5 | + * | XXH3_128bits() | @b AVX2 | 128 | 57.9 GB/s | 118.1 | + * | CLHash | PCLMUL | 64 | 37.1 GB/s | 58.1 | + * | XXH3_64bits() | @b SSE2 | 64 | 31.5 GB/s | 133.1 | + * | XXH3_128bits() | @b SSE2 | 128 | 29.6 GB/s | 118.1 | + * | RAM sequential read | | N/A | 28.0 GB/s | N/A | + * | ahash | AES-NI | 64 | 22.5 GB/s | 107.2 | + * | City64 | | 64 | 22.0 GB/s | 76.6 | + * | T1ha2 | | 64 | 22.0 GB/s | 99.0 | + * | City128 | | 128 | 21.7 GB/s | 57.7 | + * | FarmHash | AES-NI | 64 | 21.3 GB/s | 71.9 | + * | XXH64() | | 64 | 19.4 GB/s | 71.0 | + * | SpookyHash | | 64 | 19.3 GB/s | 53.2 | + * | Mum | | 64 | 18.0 GB/s | 67.0 | + * | CRC32C | SSE4.2 | 32 | 13.0 GB/s | 57.9 | + * | XXH32() | | 32 | 9.7 GB/s | 71.9 | + * | City32 | | 32 | 9.1 GB/s | 66.0 | + * | Blake3* | @b AVX2 | 256 | 4.4 GB/s | 8.1 | + * | Murmur3 | | 32 | 3.9 GB/s | 56.1 | + * | SipHash* | | 64 | 3.0 GB/s | 43.2 | + * | Blake3* | @b SSE2 | 256 | 2.4 GB/s | 8.1 | + * | HighwayHash | | 64 | 1.4 GB/s | 6.0 | + * | FNV64 | | 64 | 1.2 GB/s | 62.7 | + * | Blake2* | | 256 | 1.1 GB/s | 5.1 | + * | SHA1* | | 160 | 0.8 GB/s | 5.6 | + * | MD5* | | 128 | 0.6 GB/s | 7.8 | + * @note + * - Hashes which require a specific ISA extension are noted. SSE2 is also noted, + * even though it is mandatory on x64. + * - Hashes with an asterisk are cryptographic. Note that MD5 is non-cryptographic + * by modern standards. + * - Small data velocity is a rough average of algorithm's efficiency for small + * data. For more accurate information, see the wiki. + * - More benchmarks and strength tests are found on the wiki: + * https://github.com/Cyan4973/xxHash/wiki + * + * Usage + * ------ + * All xxHash variants use a similar API. Changing the algorithm is a trivial + * substitution. + * + * @pre + * For functions which take an input and length parameter, the following + * requirements are assumed: + * - The range from [`input`, `input + length`) is valid, readable memory. + * - The only exception is if the `length` is `0`, `input` may be `NULL`. + * - For C++, the objects must have the *TriviallyCopyable* property, as the + * functions access bytes directly as if it was an array of `unsigned char`. + * + * @anchor single_shot_example + * **Single Shot** + * + * These functions are stateless functions which hash a contiguous block of memory, + * immediately returning the result. They are the easiest and usually the fastest + * option. + * + * XXH32(), XXH64(), XXH3_64bits(), XXH3_128bits() + * + * @code{.c} + * #include + * #include "xxhash.h" + * + * // Example for a function which hashes a null terminated string with XXH32(). + * XXH32_hash_t hash_string(const char* string, XXH32_hash_t seed) + * { + * // NULL pointers are only valid if the length is zero + * size_t length = (string == NULL) ? 0 : strlen(string); + * return XXH32(string, length, seed); + * } + * @endcode + * + * + * @anchor streaming_example + * **Streaming** + * + * These groups of functions allow incremental hashing of unknown size, even + * more than what would fit in a size_t. + * + * XXH32_reset(), XXH64_reset(), XXH3_64bits_reset(), XXH3_128bits_reset() + * + * @code{.c} + * #include + * #include + * #include "xxhash.h" + * // Example for a function which hashes a FILE incrementally with XXH3_64bits(). + * XXH64_hash_t hashFile(FILE* f) + * { + * // Allocate a state struct. Do not just use malloc() or new. + * XXH3_state_t* state = XXH3_createState(); + * assert(state != NULL && "Out of memory!"); + * // Reset the state to start a new hashing session. + * XXH3_64bits_reset(state); + * char buffer[4096]; + * size_t count; + * // Read the file in chunks + * while ((count = fread(buffer, 1, sizeof(buffer), f)) != 0) { + * // Run update() as many times as necessary to process the data + * XXH3_64bits_update(state, buffer, count); + * } + * // Retrieve the finalized hash. This will not change the state. + * XXH64_hash_t result = XXH3_64bits_digest(state); + * // Free the state. Do not use free(). + * XXH3_freeState(state); + * return result; + * } + * @endcode + * + * Streaming functions generate the xxHash value from an incremental input. + * This method is slower than single-call functions, due to state management. + * For small inputs, prefer `XXH32()` and `XXH64()`, which are better optimized. + * + * An XXH state must first be allocated using `XXH*_createState()`. + * + * Start a new hash by initializing the state with a seed using `XXH*_reset()`. + * + * Then, feed the hash state by calling `XXH*_update()` as many times as necessary. + * + * The function returns an error code, with 0 meaning OK, and any other value + * meaning there is an error. + * + * Finally, a hash value can be produced anytime, by using `XXH*_digest()`. + * This function returns the nn-bits hash as an int or long long. + * + * It's still possible to continue inserting input into the hash state after a + * digest, and generate new hash values later on by invoking `XXH*_digest()`. + * + * When done, release the state using `XXH*_freeState()`. + * + * + * @anchor canonical_representation_example + * **Canonical Representation** + * + * The default return values from XXH functions are unsigned 32, 64 and 128 bit + * integers. + * This the simplest and fastest format for further post-processing. + * + * However, this leaves open the question of what is the order on the byte level, + * since little and big endian conventions will store the same number differently. + * + * The canonical representation settles this issue by mandating big-endian + * convention, the same convention as human-readable numbers (large digits first). + * + * When writing hash values to storage, sending them over a network, or printing + * them, it's highly recommended to use the canonical representation to ensure + * portability across a wider range of systems, present and future. + * + * The following functions allow transformation of hash values to and from + * canonical format. + * + * XXH32_canonicalFromHash(), XXH32_hashFromCanonical(), + * XXH64_canonicalFromHash(), XXH64_hashFromCanonical(), + * XXH128_canonicalFromHash(), XXH128_hashFromCanonical(), + * + * @code{.c} + * #include + * #include "xxhash.h" + * + * // Example for a function which prints XXH32_hash_t in human readable format + * void printXxh32(XXH32_hash_t hash) + * { + * XXH32_canonical_t cano; + * XXH32_canonicalFromHash(&cano, hash); + * size_t i; + * for(i = 0; i < sizeof(cano.digest); ++i) { + * printf("%02x", cano.digest[i]); + * } + * printf("\n"); + * } + * + * // Example for a function which converts XXH32_canonical_t to XXH32_hash_t + * XXH32_hash_t convertCanonicalToXxh32(XXH32_canonical_t cano) + * { + * XXH32_hash_t hash = XXH32_hashFromCanonical(&cano); + * return hash; + * } + * @endcode + * + * + * @file xxhash.h + * xxHash prototypes and implementation + */ + +#if defined (__cplusplus) +extern "C" { +#endif + +/* **************************** + * INLINE mode + ******************************/ +/*! + * @defgroup public Public API + * Contains details on the public xxHash functions. + * @{ + */ +#ifdef XXH_DOXYGEN +/*! + * @brief Gives access to internal state declaration, required for static allocation. + * + * Incompatible with dynamic linking, due to risks of ABI changes. + * + * Usage: + * @code{.c} + * #define XXH_STATIC_LINKING_ONLY + * #include "xxhash.h" + * @endcode + */ +# define XXH_STATIC_LINKING_ONLY +/* Do not undef XXH_STATIC_LINKING_ONLY for Doxygen */ + +/*! + * @brief Gives access to internal definitions. + * + * Usage: + * @code{.c} + * #define XXH_STATIC_LINKING_ONLY + * #define XXH_IMPLEMENTATION + * #include "xxhash.h" + * @endcode + */ +# define XXH_IMPLEMENTATION +/* Do not undef XXH_IMPLEMENTATION for Doxygen */ + +/*! + * @brief Exposes the implementation and marks all functions as `inline`. + * + * Use these build macros to inline xxhash into the target unit. + * Inlining improves performance on small inputs, especially when the length is + * expressed as a compile-time constant: + * + * https://fastcompression.blogspot.com/2018/03/xxhash-for-small-keys-impressive-power.html + * + * It also keeps xxHash symbols private to the unit, so they are not exported. + * + * Usage: + * @code{.c} + * #define XXH_INLINE_ALL + * #include "xxhash.h" + * @endcode + * Do not compile and link xxhash.o as a separate object, as it is not useful. + */ +# define XXH_INLINE_ALL +# undef XXH_INLINE_ALL +/*! + * @brief Exposes the implementation without marking functions as inline. + */ +# define XXH_PRIVATE_API +# undef XXH_PRIVATE_API +/*! + * @brief Emulate a namespace by transparently prefixing all symbols. + * + * If you want to include _and expose_ xxHash functions from within your own + * library, but also want to avoid symbol collisions with other libraries which + * may also include xxHash, you can use @ref XXH_NAMESPACE to automatically prefix + * any public symbol from xxhash library with the value of @ref XXH_NAMESPACE + * (therefore, avoid empty or numeric values). + * + * Note that no change is required within the calling program as long as it + * includes `xxhash.h`: Regular symbol names will be automatically translated + * by this header. + */ +# define XXH_NAMESPACE /* YOUR NAME HERE */ +# undef XXH_NAMESPACE +#endif + +#if (defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API)) \ + && !defined(XXH_INLINE_ALL_31684351384) + /* this section should be traversed only once */ +# define XXH_INLINE_ALL_31684351384 + /* give access to the advanced API, required to compile implementations */ +# undef XXH_STATIC_LINKING_ONLY /* avoid macro redef */ +# define XXH_STATIC_LINKING_ONLY + /* make all functions private */ +# undef XXH_PUBLIC_API +# if defined(__GNUC__) +# define XXH_PUBLIC_API static __inline __attribute__((__unused__)) +# elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) +# define XXH_PUBLIC_API static inline +# elif defined(_MSC_VER) +# define XXH_PUBLIC_API static __inline +# else + /* note: this version may generate warnings for unused static functions */ +# define XXH_PUBLIC_API static +# endif + + /* + * This part deals with the special case where a unit wants to inline xxHash, + * but "xxhash.h" has previously been included without XXH_INLINE_ALL, + * such as part of some previously included *.h header file. + * Without further action, the new include would just be ignored, + * and functions would effectively _not_ be inlined (silent failure). + * The following macros solve this situation by prefixing all inlined names, + * avoiding naming collision with previous inclusions. + */ + /* Before that, we unconditionally #undef all symbols, + * in case they were already defined with XXH_NAMESPACE. + * They will then be redefined for XXH_INLINE_ALL + */ +# undef XXH_versionNumber + /* XXH32 */ +# undef XXH32 +# undef XXH32_createState +# undef XXH32_freeState +# undef XXH32_reset +# undef XXH32_update +# undef XXH32_digest +# undef XXH32_copyState +# undef XXH32_canonicalFromHash +# undef XXH32_hashFromCanonical + /* XXH64 */ +# undef XXH64 +# undef XXH64_createState +# undef XXH64_freeState +# undef XXH64_reset +# undef XXH64_update +# undef XXH64_digest +# undef XXH64_copyState +# undef XXH64_canonicalFromHash +# undef XXH64_hashFromCanonical + /* XXH3_64bits */ +# undef XXH3_64bits +# undef XXH3_64bits_withSecret +# undef XXH3_64bits_withSeed +# undef XXH3_64bits_withSecretandSeed +# undef XXH3_createState +# undef XXH3_freeState +# undef XXH3_copyState +# undef XXH3_64bits_reset +# undef XXH3_64bits_reset_withSeed +# undef XXH3_64bits_reset_withSecret +# undef XXH3_64bits_update +# undef XXH3_64bits_digest +# undef XXH3_generateSecret + /* XXH3_128bits */ +# undef XXH128 +# undef XXH3_128bits +# undef XXH3_128bits_withSeed +# undef XXH3_128bits_withSecret +# undef XXH3_128bits_reset +# undef XXH3_128bits_reset_withSeed +# undef XXH3_128bits_reset_withSecret +# undef XXH3_128bits_reset_withSecretandSeed +# undef XXH3_128bits_update +# undef XXH3_128bits_digest +# undef XXH128_isEqual +# undef XXH128_cmp +# undef XXH128_canonicalFromHash +# undef XXH128_hashFromCanonical + /* Finally, free the namespace itself */ +# undef XXH_NAMESPACE + + /* employ the namespace for XXH_INLINE_ALL */ +# define XXH_NAMESPACE XXH_INLINE_ + /* + * Some identifiers (enums, type names) are not symbols, + * but they must nonetheless be renamed to avoid redeclaration. + * Alternative solution: do not redeclare them. + * However, this requires some #ifdefs, and has a more dispersed impact. + * Meanwhile, renaming can be achieved in a single place. + */ +# define XXH_IPREF(Id) XXH_NAMESPACE ## Id +# define XXH_OK XXH_IPREF(XXH_OK) +# define XXH_ERROR XXH_IPREF(XXH_ERROR) +# define XXH_errorcode XXH_IPREF(XXH_errorcode) +# define XXH32_canonical_t XXH_IPREF(XXH32_canonical_t) +# define XXH64_canonical_t XXH_IPREF(XXH64_canonical_t) +# define XXH128_canonical_t XXH_IPREF(XXH128_canonical_t) +# define XXH32_state_s XXH_IPREF(XXH32_state_s) +# define XXH32_state_t XXH_IPREF(XXH32_state_t) +# define XXH64_state_s XXH_IPREF(XXH64_state_s) +# define XXH64_state_t XXH_IPREF(XXH64_state_t) +# define XXH3_state_s XXH_IPREF(XXH3_state_s) +# define XXH3_state_t XXH_IPREF(XXH3_state_t) +# define XXH128_hash_t XXH_IPREF(XXH128_hash_t) + /* Ensure the header is parsed again, even if it was previously included */ +# undef XXHASH_H_5627135585666179 +# undef XXHASH_H_STATIC_13879238742 +#endif /* XXH_INLINE_ALL || XXH_PRIVATE_API */ + +/* **************************************************************** + * Stable API + *****************************************************************/ +#ifndef XXHASH_H_5627135585666179 +#define XXHASH_H_5627135585666179 1 + +/*! @brief Marks a global symbol. */ +#if !defined(XXH_INLINE_ALL) && !defined(XXH_PRIVATE_API) +# if defined(_WIN32) && defined(_MSC_VER) && (defined(XXH_IMPORT) || defined(XXH_EXPORT)) +# ifdef XXH_EXPORT +# define XXH_PUBLIC_API __declspec(dllexport) +# elif XXH_IMPORT +# define XXH_PUBLIC_API __declspec(dllimport) +# endif +# else +# define XXH_PUBLIC_API /* do nothing */ +# endif +#endif + +#ifdef XXH_NAMESPACE +# define XXH_CAT(A,B) A##B +# define XXH_NAME2(A,B) XXH_CAT(A,B) +# define XXH_versionNumber XXH_NAME2(XXH_NAMESPACE, XXH_versionNumber) +/* XXH32 */ +# define XXH32 XXH_NAME2(XXH_NAMESPACE, XXH32) +# define XXH32_createState XXH_NAME2(XXH_NAMESPACE, XXH32_createState) +# define XXH32_freeState XXH_NAME2(XXH_NAMESPACE, XXH32_freeState) +# define XXH32_reset XXH_NAME2(XXH_NAMESPACE, XXH32_reset) +# define XXH32_update XXH_NAME2(XXH_NAMESPACE, XXH32_update) +# define XXH32_digest XXH_NAME2(XXH_NAMESPACE, XXH32_digest) +# define XXH32_copyState XXH_NAME2(XXH_NAMESPACE, XXH32_copyState) +# define XXH32_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH32_canonicalFromHash) +# define XXH32_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH32_hashFromCanonical) +/* XXH64 */ +# define XXH64 XXH_NAME2(XXH_NAMESPACE, XXH64) +# define XXH64_createState XXH_NAME2(XXH_NAMESPACE, XXH64_createState) +# define XXH64_freeState XXH_NAME2(XXH_NAMESPACE, XXH64_freeState) +# define XXH64_reset XXH_NAME2(XXH_NAMESPACE, XXH64_reset) +# define XXH64_update XXH_NAME2(XXH_NAMESPACE, XXH64_update) +# define XXH64_digest XXH_NAME2(XXH_NAMESPACE, XXH64_digest) +# define XXH64_copyState XXH_NAME2(XXH_NAMESPACE, XXH64_copyState) +# define XXH64_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH64_canonicalFromHash) +# define XXH64_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH64_hashFromCanonical) +/* XXH3_64bits */ +# define XXH3_64bits XXH_NAME2(XXH_NAMESPACE, XXH3_64bits) +# define XXH3_64bits_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSecret) +# define XXH3_64bits_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSeed) +# define XXH3_64bits_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSecretandSeed) +# define XXH3_createState XXH_NAME2(XXH_NAMESPACE, XXH3_createState) +# define XXH3_freeState XXH_NAME2(XXH_NAMESPACE, XXH3_freeState) +# define XXH3_copyState XXH_NAME2(XXH_NAMESPACE, XXH3_copyState) +# define XXH3_64bits_reset XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset) +# define XXH3_64bits_reset_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSeed) +# define XXH3_64bits_reset_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSecret) +# define XXH3_64bits_reset_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSecretandSeed) +# define XXH3_64bits_update XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_update) +# define XXH3_64bits_digest XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_digest) +# define XXH3_generateSecret XXH_NAME2(XXH_NAMESPACE, XXH3_generateSecret) +# define XXH3_generateSecret_fromSeed XXH_NAME2(XXH_NAMESPACE, XXH3_generateSecret_fromSeed) +/* XXH3_128bits */ +# define XXH128 XXH_NAME2(XXH_NAMESPACE, XXH128) +# define XXH3_128bits XXH_NAME2(XXH_NAMESPACE, XXH3_128bits) +# define XXH3_128bits_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSeed) +# define XXH3_128bits_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSecret) +# define XXH3_128bits_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSecretandSeed) +# define XXH3_128bits_reset XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset) +# define XXH3_128bits_reset_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSeed) +# define XXH3_128bits_reset_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSecret) +# define XXH3_128bits_reset_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSecretandSeed) +# define XXH3_128bits_update XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_update) +# define XXH3_128bits_digest XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_digest) +# define XXH128_isEqual XXH_NAME2(XXH_NAMESPACE, XXH128_isEqual) +# define XXH128_cmp XXH_NAME2(XXH_NAMESPACE, XXH128_cmp) +# define XXH128_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH128_canonicalFromHash) +# define XXH128_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH128_hashFromCanonical) +#endif + + +/* ************************************* +* Compiler specifics +***************************************/ + +/* specific declaration modes for Windows */ +#if !defined(XXH_INLINE_ALL) && !defined(XXH_PRIVATE_API) +# if defined(_WIN32) && defined(_MSC_VER) && (defined(XXH_IMPORT) || defined(XXH_EXPORT)) +# ifdef XXH_EXPORT +# define XXH_PUBLIC_API __declspec(dllexport) +# elif XXH_IMPORT +# define XXH_PUBLIC_API __declspec(dllimport) +# endif +# else +# define XXH_PUBLIC_API /* do nothing */ +# endif +#endif + +#if defined (__GNUC__) +# define XXH_CONSTF __attribute__((__const__)) +# define XXH_PUREF __attribute__((__pure__)) +# define XXH_MALLOCF __attribute__((__malloc__)) +#else +# define XXH_CONSTF /* disable */ +# define XXH_PUREF +# define XXH_MALLOCF +#endif + +/* ************************************* +* Version +***************************************/ +#define XXH_VERSION_MAJOR 0 +#define XXH_VERSION_MINOR 8 +#define XXH_VERSION_RELEASE 3 +/*! @brief Version number, encoded as two digits each */ +#define XXH_VERSION_NUMBER (XXH_VERSION_MAJOR *100*100 + XXH_VERSION_MINOR *100 + XXH_VERSION_RELEASE) + +/*! + * @brief Obtains the xxHash version. + * + * This is mostly useful when xxHash is compiled as a shared library, + * since the returned value comes from the library, as opposed to header file. + * + * @return @ref XXH_VERSION_NUMBER of the invoked library. + */ +XXH_PUBLIC_API XXH_CONSTF unsigned XXH_versionNumber (void); + + +/* **************************** +* Common basic types +******************************/ +#include /* size_t */ +/*! + * @brief Exit code for the streaming API. + */ +typedef enum { + XXH_OK = 0, /*!< OK */ + XXH_ERROR /*!< Error */ +} XXH_errorcode; + + +/*-********************************************************************** +* 32-bit hash +************************************************************************/ +#if defined(XXH_DOXYGEN) /* Don't show include */ +/*! + * @brief An unsigned 32-bit integer. + * + * Not necessarily defined to `uint32_t` but functionally equivalent. + */ +typedef uint32_t XXH32_hash_t; + +#elif !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) +# ifdef _AIX +# include +# else +# include +# endif + typedef uint32_t XXH32_hash_t; + +#else +# include +# if UINT_MAX == 0xFFFFFFFFUL + typedef unsigned int XXH32_hash_t; +# elif ULONG_MAX == 0xFFFFFFFFUL + typedef unsigned long XXH32_hash_t; +# else +# error "unsupported platform: need a 32-bit type" +# endif +#endif + +/*! + * @} + * + * @defgroup XXH32_family XXH32 family + * @ingroup public + * Contains functions used in the classic 32-bit xxHash algorithm. + * + * @note + * XXH32 is useful for older platforms, with no or poor 64-bit performance. + * Note that the @ref XXH3_family provides competitive speed for both 32-bit + * and 64-bit systems, and offers true 64/128 bit hash results. + * + * @see @ref XXH64_family, @ref XXH3_family : Other xxHash families + * @see @ref XXH32_impl for implementation details + * @{ + */ + +/*! + * @brief Calculates the 32-bit hash of @p input using xxHash32. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * @param seed The 32-bit seed to alter the hash's output predictably. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 32-bit xxHash32 value. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32 (const void* input, size_t length, XXH32_hash_t seed); + +#ifndef XXH_NO_STREAM +/*! + * @typedef struct XXH32_state_s XXH32_state_t + * @brief The opaque state struct for the XXH32 streaming API. + * + * @see XXH32_state_s for details. + * @see @ref streaming_example "Streaming Example" + */ +typedef struct XXH32_state_s XXH32_state_t; + +/*! + * @brief Allocates an @ref XXH32_state_t. + * + * @return An allocated pointer of @ref XXH32_state_t on success. + * @return `NULL` on failure. + * + * @note Must be freed with XXH32_freeState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_MALLOCF XXH32_state_t* XXH32_createState(void); +/*! + * @brief Frees an @ref XXH32_state_t. + * + * @param statePtr A pointer to an @ref XXH32_state_t allocated with @ref XXH32_createState(). + * + * @return @ref XXH_OK. + * + * @note @p statePtr must be allocated with XXH32_createState(). + * + * @see @ref streaming_example "Streaming Example" + * + */ +XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr); +/*! + * @brief Copies one @ref XXH32_state_t to another. + * + * @param dst_state The state to copy to. + * @param src_state The state to copy from. + * @pre + * @p dst_state and @p src_state must not be `NULL` and must not overlap. + */ +XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dst_state, const XXH32_state_t* src_state); + +/*! + * @brief Resets an @ref XXH32_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 32-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note This function resets and seeds a state. Call it before @ref XXH32_update(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH32_reset (XXH32_state_t* statePtr, XXH32_hash_t seed); + +/*! + * @brief Consumes a block of @p input to an @ref XXH32_state_t. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note Call this to incrementally consume blocks of data. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* statePtr, const void* input, size_t length); + +/*! + * @brief Returns the calculated hash value from an @ref XXH32_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated 32-bit xxHash32 value from that state. + * + * @note + * Calling XXH32_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32_digest (const XXH32_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ + +/******* Canonical representation *******/ + +/*! + * @brief Canonical (big endian) representation of @ref XXH32_hash_t. + */ +typedef struct { + unsigned char digest[4]; /*!< Hash bytes, big endian */ +} XXH32_canonical_t; + +/*! + * @brief Converts an @ref XXH32_hash_t to a big endian @ref XXH32_canonical_t. + * + * @param dst The @ref XXH32_canonical_t pointer to be stored to. + * @param hash The @ref XXH32_hash_t to be converted. + * + * @pre + * @p dst must not be `NULL`. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash); + +/*! + * @brief Converts an @ref XXH32_canonical_t to a native @ref XXH32_hash_t. + * + * @param src The @ref XXH32_canonical_t to convert. + * + * @pre + * @p src must not be `NULL`. + * + * @return The converted hash. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src); + + +/*! @cond Doxygen ignores this part */ +#ifdef __has_attribute +# define XXH_HAS_ATTRIBUTE(x) __has_attribute(x) +#else +# define XXH_HAS_ATTRIBUTE(x) 0 +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* + * C23 __STDC_VERSION__ number hasn't been specified yet. For now + * leave as `201711L` (C17 + 1). + * TODO: Update to correct value when its been specified. + */ +#define XXH_C23_VN 201711L +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* C-language Attributes are added in C23. */ +#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= XXH_C23_VN) && defined(__has_c_attribute) +# define XXH_HAS_C_ATTRIBUTE(x) __has_c_attribute(x) +#else +# define XXH_HAS_C_ATTRIBUTE(x) 0 +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +#if defined(__cplusplus) && defined(__has_cpp_attribute) +# define XXH_HAS_CPP_ATTRIBUTE(x) __has_cpp_attribute(x) +#else +# define XXH_HAS_CPP_ATTRIBUTE(x) 0 +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* + * Define XXH_FALLTHROUGH macro for annotating switch case with the 'fallthrough' attribute + * introduced in CPP17 and C23. + * CPP17 : https://en.cppreference.com/w/cpp/language/attributes/fallthrough + * C23 : https://en.cppreference.com/w/c/language/attributes/fallthrough + */ +#if XXH_HAS_C_ATTRIBUTE(fallthrough) || XXH_HAS_CPP_ATTRIBUTE(fallthrough) +# define XXH_FALLTHROUGH [[fallthrough]] +#elif XXH_HAS_ATTRIBUTE(__fallthrough__) +# define XXH_FALLTHROUGH __attribute__ ((__fallthrough__)) +#else +# define XXH_FALLTHROUGH /* fallthrough */ +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* + * Define XXH_NOESCAPE for annotated pointers in public API. + * https://clang.llvm.org/docs/AttributeReference.html#noescape + * As of writing this, only supported by clang. + */ +#if XXH_HAS_ATTRIBUTE(noescape) +# define XXH_NOESCAPE __attribute__((__noescape__)) +#else +# define XXH_NOESCAPE +#endif +/*! @endcond */ + + +/*! + * @} + * @ingroup public + * @{ + */ + +#ifndef XXH_NO_LONG_LONG +/*-********************************************************************** +* 64-bit hash +************************************************************************/ +#if defined(XXH_DOXYGEN) /* don't include */ +/*! + * @brief An unsigned 64-bit integer. + * + * Not necessarily defined to `uint64_t` but functionally equivalent. + */ +typedef uint64_t XXH64_hash_t; +#elif !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) +# ifdef _AIX +# include +# else +# include +# endif + typedef uint64_t XXH64_hash_t; +#else +# include +# if defined(__LP64__) && ULONG_MAX == 0xFFFFFFFFFFFFFFFFULL + /* LP64 ABI says uint64_t is unsigned long */ + typedef unsigned long XXH64_hash_t; +# else + /* the following type must have a width of 64-bit */ + typedef unsigned long long XXH64_hash_t; +# endif +#endif + +/*! + * @} + * + * @defgroup XXH64_family XXH64 family + * @ingroup public + * @{ + * Contains functions used in the classic 64-bit xxHash algorithm. + * + * @note + * XXH3 provides competitive speed for both 32-bit and 64-bit systems, + * and offers true 64/128 bit hash results. + * It provides better speed for systems with vector processing capabilities. + */ + +/*! + * @brief Calculates the 64-bit hash of @p input using xxHash64. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * @param seed The 64-bit seed to alter the hash's output predictably. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 64-bit xxHash64 value. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed); + +/******* Streaming *******/ +#ifndef XXH_NO_STREAM +/*! + * @brief The opaque state struct for the XXH64 streaming API. + * + * @see XXH64_state_s for details. + * @see @ref streaming_example "Streaming Example" + */ +typedef struct XXH64_state_s XXH64_state_t; /* incomplete type */ + +/*! + * @brief Allocates an @ref XXH64_state_t. + * + * @return An allocated pointer of @ref XXH64_state_t on success. + * @return `NULL` on failure. + * + * @note Must be freed with XXH64_freeState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_MALLOCF XXH64_state_t* XXH64_createState(void); + +/*! + * @brief Frees an @ref XXH64_state_t. + * + * @param statePtr A pointer to an @ref XXH64_state_t allocated with @ref XXH64_createState(). + * + * @return @ref XXH_OK. + * + * @note @p statePtr must be allocated with XXH64_createState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr); + +/*! + * @brief Copies one @ref XXH64_state_t to another. + * + * @param dst_state The state to copy to. + * @param src_state The state to copy from. + * @pre + * @p dst_state and @p src_state must not be `NULL` and must not overlap. + */ +XXH_PUBLIC_API void XXH64_copyState(XXH_NOESCAPE XXH64_state_t* dst_state, const XXH64_state_t* src_state); + +/*! + * @brief Resets an @ref XXH64_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note This function resets and seeds a state. Call it before @ref XXH64_update(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH64_reset (XXH_NOESCAPE XXH64_state_t* statePtr, XXH64_hash_t seed); + +/*! + * @brief Consumes a block of @p input to an @ref XXH64_state_t. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note Call this to incrementally consume blocks of data. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH_NOESCAPE XXH64_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Returns the calculated hash value from an @ref XXH64_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated 64-bit xxHash64 value from that state. + * + * @note + * Calling XXH64_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64_digest (XXH_NOESCAPE const XXH64_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ +/******* Canonical representation *******/ + +/*! + * @brief Canonical (big endian) representation of @ref XXH64_hash_t. + */ +typedef struct { unsigned char digest[sizeof(XXH64_hash_t)]; } XXH64_canonical_t; + +/*! + * @brief Converts an @ref XXH64_hash_t to a big endian @ref XXH64_canonical_t. + * + * @param dst The @ref XXH64_canonical_t pointer to be stored to. + * @param hash The @ref XXH64_hash_t to be converted. + * + * @pre + * @p dst must not be `NULL`. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH_NOESCAPE XXH64_canonical_t* dst, XXH64_hash_t hash); + +/*! + * @brief Converts an @ref XXH64_canonical_t to a native @ref XXH64_hash_t. + * + * @param src The @ref XXH64_canonical_t to convert. + * + * @pre + * @p src must not be `NULL`. + * + * @return The converted hash. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64_hashFromCanonical(XXH_NOESCAPE const XXH64_canonical_t* src); + +#ifndef XXH_NO_XXH3 + +/*! + * @} + * ************************************************************************ + * @defgroup XXH3_family XXH3 family + * @ingroup public + * @{ + * + * XXH3 is a more recent hash algorithm featuring: + * - Improved speed for both small and large inputs + * - True 64-bit and 128-bit outputs + * - SIMD acceleration + * - Improved 32-bit viability + * + * Speed analysis methodology is explained here: + * + * https://fastcompression.blogspot.com/2019/03/presenting-xxh3.html + * + * Compared to XXH64, expect XXH3 to run approximately + * ~2x faster on large inputs and >3x faster on small ones, + * exact differences vary depending on platform. + * + * XXH3's speed benefits greatly from SIMD and 64-bit arithmetic, + * but does not require it. + * Most 32-bit and 64-bit targets that can run XXH32 smoothly can run XXH3 + * at competitive speeds, even without vector support. Further details are + * explained in the implementation. + * + * XXH3 has a fast scalar implementation, but it also includes accelerated SIMD + * implementations for many common platforms: + * - AVX512 + * - AVX2 + * - SSE2 + * - ARM NEON + * - WebAssembly SIMD128 + * - POWER8 VSX + * - s390x ZVector + * This can be controlled via the @ref XXH_VECTOR macro, but it automatically + * selects the best version according to predefined macros. For the x86 family, an + * automatic runtime dispatcher is included separately in @ref xxh_x86dispatch.c. + * + * XXH3 implementation is portable: + * it has a generic C90 formulation that can be compiled on any platform, + * all implementations generate exactly the same hash value on all platforms. + * Starting from v0.8.0, it's also labelled "stable", meaning that + * any future version will also generate the same hash value. + * + * XXH3 offers 2 variants, _64bits and _128bits. + * + * When only 64 bits are needed, prefer invoking the _64bits variant, as it + * reduces the amount of mixing, resulting in faster speed on small inputs. + * It's also generally simpler to manipulate a scalar return type than a struct. + * + * The API supports one-shot hashing, streaming mode, and custom secrets. + */ +/*-********************************************************************** +* XXH3 64-bit variant +************************************************************************/ + +/*! + * @brief Calculates 64-bit unseeded variant of XXH3 hash of @p input. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 64-bit XXH3 hash value. + * + * @note + * This is equivalent to @ref XXH3_64bits_withSeed() with a seed of `0`, however + * it may have slightly better performance due to constant propagation of the + * defaults. + * + * @see + * XXH3_64bits_withSeed(), XXH3_64bits_withSecret(): other seeding variants + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits(XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Calculates 64-bit seeded variant of XXH3 hash of @p input. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 64-bit XXH3 hash value. + * + * @note + * seed == 0 produces the same results as @ref XXH3_64bits(). + * + * This variant generates a custom secret on the fly based on default secret + * altered using the @p seed value. + * + * While this operation is decently fast, note that it's not completely free. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_withSeed(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed); + +/*! + * The bare minimum size for a custom secret. + * + * @see + * XXH3_64bits_withSecret(), XXH3_64bits_reset_withSecret(), + * XXH3_128bits_withSecret(), XXH3_128bits_reset_withSecret(). + */ +#define XXH3_SECRET_SIZE_MIN 136 + +/*! + * @brief Calculates 64-bit variant of XXH3 with a custom "secret". + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @return The calculated 64-bit XXH3 hash value. + * + * @pre + * The memory between @p data and @p data + @p len must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p data may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * It's possible to provide any blob of bytes as a "secret" to generate the hash. + * This makes it more difficult for an external actor to prepare an intentional collision. + * The main condition is that @p secretSize *must* be large enough (>= @ref XXH3_SECRET_SIZE_MIN). + * However, the quality of the secret impacts the dispersion of the hash algorithm. + * Therefore, the secret _must_ look like a bunch of random bytes. + * Avoid "trivial" or structured data such as repeated sequences or a text document. + * Whenever in doubt about the "randomness" of the blob of bytes, + * consider employing @ref XXH3_generateSecret() instead (see below). + * It will generate a proper high entropy secret derived from the blob of bytes. + * Another advantage of using XXH3_generateSecret() is that + * it guarantees that all bits within the initial blob of bytes + * will impact every bit of the output. + * This is not necessarily the case when using the blob of bytes directly + * because, when hashing _small_ inputs, only a portion of the secret is employed. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_withSecret(XXH_NOESCAPE const void* data, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize); + + +/******* Streaming *******/ +#ifndef XXH_NO_STREAM +/* + * Streaming requires state maintenance. + * This operation costs memory and CPU. + * As a consequence, streaming is slower than one-shot hashing. + * For better performance, prefer one-shot functions whenever applicable. + */ + +/*! + * @brief The opaque state struct for the XXH3 streaming API. + * + * @see XXH3_state_s for details. + * @see @ref streaming_example "Streaming Example" + */ +typedef struct XXH3_state_s XXH3_state_t; +XXH_PUBLIC_API XXH_MALLOCF XXH3_state_t* XXH3_createState(void); +XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr); + +/*! + * @brief Copies one @ref XXH3_state_t to another. + * + * @param dst_state The state to copy to. + * @param src_state The state to copy from. + * @pre + * @p dst_state and @p src_state must not be `NULL` and must not overlap. + */ +XXH_PUBLIC_API void XXH3_copyState(XXH_NOESCAPE XXH3_state_t* dst_state, XXH_NOESCAPE const XXH3_state_t* src_state); + +/*! + * @brief Resets an @ref XXH3_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret with default parameters. + * - Call this function before @ref XXH3_64bits_update(). + * - Digest will be equivalent to `XXH3_64bits()`. + * + * @see @ref streaming_example "Streaming Example" + * + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr); + +/*! + * @brief Resets an @ref XXH3_state_t with 64-bit seed to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret from `seed`. + * - Call this function before @ref XXH3_64bits_update(). + * - Digest will be equivalent to `XXH3_64bits_withSeed()`. + * + * @see @ref streaming_example "Streaming Example" + * + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed); + +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * `secret` is referenced, it _must outlive_ the hash streaming session. + * + * Similar to one-shot API, `secretSize` must be >= @ref XXH3_SECRET_SIZE_MIN, + * and the quality of produced hash values depends on secret's entropy + * (secret's content should look like a bunch of random bytes). + * When in doubt about the randomness of a candidate `secret`, + * consider employing `XXH3_generateSecret()` instead (see below). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize); + +/*! + * @brief Consumes a block of @p input to an @ref XXH3_state_t. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note Call this to incrementally consume blocks of data. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update (XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Returns the calculated XXH3 64-bit hash value from an @ref XXH3_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated XXH3 64-bit hash value from that state. + * + * @note + * Calling XXH3_64bits_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_digest (XXH_NOESCAPE const XXH3_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ + +/* note : canonical representation of XXH3 is the same as XXH64 + * since they both produce XXH64_hash_t values */ + + +/*-********************************************************************** +* XXH3 128-bit variant +************************************************************************/ + +/*! + * @brief The return value from 128-bit hashes. + * + * Stored in little endian order, although the fields themselves are in native + * endianness. + */ +typedef struct { + XXH64_hash_t low64; /*!< `value & 0xFFFFFFFFFFFFFFFF` */ + XXH64_hash_t high64; /*!< `value >> 64` */ +} XXH128_hash_t; + +/*! + * @brief Calculates 128-bit unseeded variant of XXH3 of @p data. + * + * @param data The block of data to be hashed, at least @p length bytes in size. + * @param len The length of @p data, in bytes. + * + * @return The calculated 128-bit variant of XXH3 value. + * + * The 128-bit variant of XXH3 has more strength, but it has a bit of overhead + * for shorter inputs. + * + * This is equivalent to @ref XXH3_128bits_withSeed() with a seed of `0`, however + * it may have slightly better performance due to constant propagation of the + * defaults. + * + * @see XXH3_128bits_withSeed(), XXH3_128bits_withSecret(): other seeding variants + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits(XXH_NOESCAPE const void* data, size_t len); +/*! @brief Calculates 128-bit seeded variant of XXH3 hash of @p data. + * + * @param data The block of data to be hashed, at least @p length bytes in size. + * @param len The length of @p data, in bytes. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @return The calculated 128-bit variant of XXH3 value. + * + * @note + * seed == 0 produces the same results as @ref XXH3_64bits(). + * + * This variant generates a custom secret on the fly based on default secret + * altered using the @p seed value. + * + * While this operation is decently fast, note that it's not completely free. + * + * @see XXH3_128bits(), XXH3_128bits_withSecret(): other seeding variants + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_withSeed(XXH_NOESCAPE const void* data, size_t len, XXH64_hash_t seed); +/*! + * @brief Calculates 128-bit variant of XXH3 with a custom "secret". + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @return The calculated 128-bit variant of XXH3 value. + * + * It's possible to provide any blob of bytes as a "secret" to generate the hash. + * This makes it more difficult for an external actor to prepare an intentional collision. + * The main condition is that @p secretSize *must* be large enough (>= @ref XXH3_SECRET_SIZE_MIN). + * However, the quality of the secret impacts the dispersion of the hash algorithm. + * Therefore, the secret _must_ look like a bunch of random bytes. + * Avoid "trivial" or structured data such as repeated sequences or a text document. + * Whenever in doubt about the "randomness" of the blob of bytes, + * consider employing @ref XXH3_generateSecret() instead (see below). + * It will generate a proper high entropy secret derived from the blob of bytes. + * Another advantage of using XXH3_generateSecret() is that + * it guarantees that all bits within the initial blob of bytes + * will impact every bit of the output. + * This is not necessarily the case when using the blob of bytes directly + * because, when hashing _small_ inputs, only a portion of the secret is employed. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_withSecret(XXH_NOESCAPE const void* data, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize); + +/******* Streaming *******/ +#ifndef XXH_NO_STREAM +/* + * Streaming requires state maintenance. + * This operation costs memory and CPU. + * As a consequence, streaming is slower than one-shot hashing. + * For better performance, prefer one-shot functions whenever applicable. + * + * XXH3_128bits uses the same XXH3_state_t as XXH3_64bits(). + * Use already declared XXH3_createState() and XXH3_freeState(). + * + * All reset and streaming functions have same meaning as their 64-bit counterpart. + */ + +/*! + * @brief Resets an @ref XXH3_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret with default parameters. + * - Call it before @ref XXH3_128bits_update(). + * - Digest will be equivalent to `XXH3_128bits()`. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr); + +/*! + * @brief Resets an @ref XXH3_state_t with 64-bit seed to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret from `seed`. + * - Call it before @ref XXH3_128bits_update(). + * - Digest will be equivalent to `XXH3_128bits_withSeed()`. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed); +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * `secret` is referenced, it _must outlive_ the hash streaming session. + * Similar to one-shot API, `secretSize` must be >= @ref XXH3_SECRET_SIZE_MIN, + * and the quality of produced hash values depends on secret's entropy + * (secret's content should look like a bunch of random bytes). + * When in doubt about the randomness of a candidate `secret`, + * consider employing `XXH3_generateSecret()` instead (see below). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize); + +/*! + * @brief Consumes a block of @p input to an @ref XXH3_state_t. + * + * Call this to incrementally consume blocks of data. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_update (XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Returns the calculated XXH3 128-bit hash value from an @ref XXH3_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated XXH3 128-bit hash value from that state. + * + * @note + * Calling XXH3_128bits_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_digest (XXH_NOESCAPE const XXH3_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ + +/* Following helper functions make it possible to compare XXH128_hast_t values. + * Since XXH128_hash_t is a structure, this capability is not offered by the language. + * Note: For better performance, these functions can be inlined using XXH_INLINE_ALL */ + +/*! + * @brief Check equality of two XXH128_hash_t values + * + * @param h1 The 128-bit hash value. + * @param h2 Another 128-bit hash value. + * + * @return `1` if `h1` and `h2` are equal. + * @return `0` if they are not. + */ +XXH_PUBLIC_API XXH_PUREF int XXH128_isEqual(XXH128_hash_t h1, XXH128_hash_t h2); + +/*! + * @brief Compares two @ref XXH128_hash_t + * + * This comparator is compatible with stdlib's `qsort()`/`bsearch()`. + * + * @param h128_1 Left-hand side value + * @param h128_2 Right-hand side value + * + * @return >0 if @p h128_1 > @p h128_2 + * @return =0 if @p h128_1 == @p h128_2 + * @return <0 if @p h128_1 < @p h128_2 + */ +XXH_PUBLIC_API XXH_PUREF int XXH128_cmp(XXH_NOESCAPE const void* h128_1, XXH_NOESCAPE const void* h128_2); + + +/******* Canonical representation *******/ +typedef struct { unsigned char digest[sizeof(XXH128_hash_t)]; } XXH128_canonical_t; + + +/*! + * @brief Converts an @ref XXH128_hash_t to a big endian @ref XXH128_canonical_t. + * + * @param dst The @ref XXH128_canonical_t pointer to be stored to. + * @param hash The @ref XXH128_hash_t to be converted. + * + * @pre + * @p dst must not be `NULL`. + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API void XXH128_canonicalFromHash(XXH_NOESCAPE XXH128_canonical_t* dst, XXH128_hash_t hash); + +/*! + * @brief Converts an @ref XXH128_canonical_t to a native @ref XXH128_hash_t. + * + * @param src The @ref XXH128_canonical_t to convert. + * + * @pre + * @p src must not be `NULL`. + * + * @return The converted hash. + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH128_hashFromCanonical(XXH_NOESCAPE const XXH128_canonical_t* src); + + +#endif /* !XXH_NO_XXH3 */ +#endif /* XXH_NO_LONG_LONG */ + +/*! + * @} + */ +#endif /* XXHASH_H_5627135585666179 */ + + + +#if defined(XXH_STATIC_LINKING_ONLY) && !defined(XXHASH_H_STATIC_13879238742) +#define XXHASH_H_STATIC_13879238742 +/* **************************************************************************** + * This section contains declarations which are not guaranteed to remain stable. + * They may change in future versions, becoming incompatible with a different + * version of the library. + * These declarations should only be used with static linking. + * Never use them in association with dynamic linking! + ***************************************************************************** */ + +/* + * These definitions are only present to allow static allocation + * of XXH states, on stack or in a struct, for example. + * Never **ever** access their members directly. + */ + +/*! + * @internal + * @brief Structure for XXH32 streaming API. + * + * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, + * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. Otherwise it is + * an opaque type. This allows fields to safely be changed. + * + * Typedef'd to @ref XXH32_state_t. + * Do not access the members of this struct directly. + * @see XXH64_state_s, XXH3_state_s + */ +struct XXH32_state_s { + XXH32_hash_t total_len_32; /*!< Total length hashed, modulo 2^32 */ + XXH32_hash_t large_len; /*!< Whether the hash is >= 16 (handles @ref total_len_32 overflow) */ + XXH32_hash_t v[4]; /*!< Accumulator lanes */ + XXH32_hash_t mem32[4]; /*!< Internal buffer for partial reads. Treated as unsigned char[16]. */ + XXH32_hash_t memsize; /*!< Amount of data in @ref mem32 */ + XXH32_hash_t reserved; /*!< Reserved field. Do not read nor write to it. */ +}; /* typedef'd to XXH32_state_t */ + + +#ifndef XXH_NO_LONG_LONG /* defined when there is no 64-bit support */ + +/*! + * @internal + * @brief Structure for XXH64 streaming API. + * + * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, + * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. Otherwise it is + * an opaque type. This allows fields to safely be changed. + * + * Typedef'd to @ref XXH64_state_t. + * Do not access the members of this struct directly. + * @see XXH32_state_s, XXH3_state_s + */ +struct XXH64_state_s { + XXH64_hash_t total_len; /*!< Total length hashed. This is always 64-bit. */ + XXH64_hash_t v[4]; /*!< Accumulator lanes */ + XXH64_hash_t mem64[4]; /*!< Internal buffer for partial reads. Treated as unsigned char[32]. */ + XXH32_hash_t memsize; /*!< Amount of data in @ref mem64 */ + XXH32_hash_t reserved32; /*!< Reserved field, needed for padding anyways*/ + XXH64_hash_t reserved64; /*!< Reserved field. Do not read or write to it. */ +}; /* typedef'd to XXH64_state_t */ + +#ifndef XXH_NO_XXH3 + +/* Windows SDK under 10.0.22000 is missing stdalign.h so we add a check + before allowing the windows compiler to use the C11 form. + Reference: https://github.com/Cyan4973/xxHash/issues/955 */ +#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) \ + && (defined(_MSC_VER) && (_MSC_VER >= 1000) || !defined(_MSC_VER)) /* >= C11 */ +# include +# define XXH_ALIGN(n) alignas(n) +#elif defined(__cplusplus) && (__cplusplus >= 201103L) /* >= C++11 */ +/* In C++ alignas() is a keyword */ +# define XXH_ALIGN(n) alignas(n) +#elif defined(__GNUC__) +# define XXH_ALIGN(n) __attribute__ ((aligned(n))) +#elif defined(_MSC_VER) +# define XXH_ALIGN(n) __declspec(align(n)) +#else +# define XXH_ALIGN(n) /* disabled */ +#endif + +/* Old GCC versions only accept the attribute after the type in structures. */ +#if !(defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L)) /* C11+ */ \ + && ! (defined(__cplusplus) && (__cplusplus >= 201103L)) /* >= C++11 */ \ + && defined(__GNUC__) +# define XXH_ALIGN_MEMBER(align, type) type XXH_ALIGN(align) +#else +# define XXH_ALIGN_MEMBER(align, type) XXH_ALIGN(align) type +#endif + +/*! + * @brief The size of the internal XXH3 buffer. + * + * This is the optimal update size for incremental hashing. + * + * @see XXH3_64b_update(), XXH3_128b_update(). + */ +#define XXH3_INTERNALBUFFER_SIZE 256 + +/*! + * @internal + * @brief Default size of the secret buffer (and @ref XXH3_kSecret). + * + * This is the size used in @ref XXH3_kSecret and the seeded functions. + * + * Not to be confused with @ref XXH3_SECRET_SIZE_MIN. + */ +#define XXH3_SECRET_DEFAULT_SIZE 192 + +/*! + * @internal + * @brief Structure for XXH3 streaming API. + * + * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, + * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. + * Otherwise it is an opaque type. + * Never use this definition in combination with dynamic library. + * This allows fields to safely be changed in the future. + * + * @note ** This structure has a strict alignment requirement of 64 bytes!! ** + * Do not allocate this with `malloc()` or `new`, + * it will not be sufficiently aligned. + * Use @ref XXH3_createState() and @ref XXH3_freeState(), or stack allocation. + * + * Typedef'd to @ref XXH3_state_t. + * Do never access the members of this struct directly. + * + * @see XXH3_INITSTATE() for stack initialization. + * @see XXH3_createState(), XXH3_freeState(). + * @see XXH32_state_s, XXH64_state_s + */ +struct XXH3_state_s { + XXH_ALIGN_MEMBER(64, XXH64_hash_t acc[8]); + /*!< The 8 accumulators. See @ref XXH32_state_s::v and @ref XXH64_state_s::v */ + XXH_ALIGN_MEMBER(64, unsigned char customSecret[XXH3_SECRET_DEFAULT_SIZE]); + /*!< Used to store a custom secret generated from a seed. */ + XXH_ALIGN_MEMBER(64, unsigned char buffer[XXH3_INTERNALBUFFER_SIZE]); + /*!< The internal buffer. @see XXH32_state_s::mem32 */ + XXH32_hash_t bufferedSize; + /*!< The amount of memory in @ref buffer, @see XXH32_state_s::memsize */ + XXH32_hash_t useSeed; + /*!< Reserved field. Needed for padding on 64-bit. */ + size_t nbStripesSoFar; + /*!< Number or stripes processed. */ + XXH64_hash_t totalLen; + /*!< Total length hashed. 64-bit even on 32-bit targets. */ + size_t nbStripesPerBlock; + /*!< Number of stripes per block. */ + size_t secretLimit; + /*!< Size of @ref customSecret or @ref extSecret */ + XXH64_hash_t seed; + /*!< Seed for _withSeed variants. Must be zero otherwise, @see XXH3_INITSTATE() */ + XXH64_hash_t reserved64; + /*!< Reserved field. */ + const unsigned char* extSecret; + /*!< Reference to an external secret for the _withSecret variants, NULL + * for other variants. */ + /* note: there may be some padding at the end due to alignment on 64 bytes */ +}; /* typedef'd to XXH3_state_t */ + +#undef XXH_ALIGN_MEMBER + +/*! + * @brief Initializes a stack-allocated `XXH3_state_s`. + * + * When the @ref XXH3_state_t structure is merely emplaced on stack, + * it should be initialized with XXH3_INITSTATE() or a memset() + * in case its first reset uses XXH3_NNbits_reset_withSeed(). + * This init can be omitted if the first reset uses default or _withSecret mode. + * This operation isn't necessary when the state is created with XXH3_createState(). + * Note that this doesn't prepare the state for a streaming operation, + * it's still necessary to use XXH3_NNbits_reset*() afterwards. + */ +#define XXH3_INITSTATE(XXH3_state_ptr) \ + do { \ + XXH3_state_t* tmp_xxh3_state_ptr = (XXH3_state_ptr); \ + tmp_xxh3_state_ptr->seed = 0; \ + tmp_xxh3_state_ptr->extSecret = NULL; \ + } while(0) + + +/*! + * @brief Calculates the 128-bit hash of @p data using XXH3. + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param seed The 64-bit seed to alter the hash's output predictably. + * + * @pre + * The memory between @p data and @p data + @p len must be valid, + * readable, contiguous memory. However, if @p len is `0`, @p data may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 128-bit XXH3 value. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH128(XXH_NOESCAPE const void* data, size_t len, XXH64_hash_t seed); + + +/* === Experimental API === */ +/* Symbols defined below must be considered tied to a specific library version. */ + +/*! + * @brief Derive a high-entropy secret from any user-defined content, named customSeed. + * + * @param secretBuffer A writable buffer for derived high-entropy secret data. + * @param secretSize Size of secretBuffer, in bytes. Must be >= XXH3_SECRET_SIZE_MIN. + * @param customSeed A user-defined content. + * @param customSeedSize Size of customSeed, in bytes. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * The generated secret can be used in combination with `*_withSecret()` functions. + * The `_withSecret()` variants are useful to provide a higher level of protection + * than 64-bit seed, as it becomes much more difficult for an external actor to + * guess how to impact the calculation logic. + * + * The function accepts as input a custom seed of any length and any content, + * and derives from it a high-entropy secret of length @p secretSize into an + * already allocated buffer @p secretBuffer. + * + * The generated secret can then be used with any `*_withSecret()` variant. + * The functions @ref XXH3_128bits_withSecret(), @ref XXH3_64bits_withSecret(), + * @ref XXH3_128bits_reset_withSecret() and @ref XXH3_64bits_reset_withSecret() + * are part of this list. They all accept a `secret` parameter + * which must be large enough for implementation reasons (>= @ref XXH3_SECRET_SIZE_MIN) + * _and_ feature very high entropy (consist of random-looking bytes). + * These conditions can be a high bar to meet, so @ref XXH3_generateSecret() can + * be employed to ensure proper quality. + * + * @p customSeed can be anything. It can have any size, even small ones, + * and its content can be anything, even "poor entropy" sources such as a bunch + * of zeroes. The resulting `secret` will nonetheless provide all required qualities. + * + * @pre + * - @p secretSize must be >= @ref XXH3_SECRET_SIZE_MIN + * - When @p customSeedSize > 0, supplying NULL as customSeed is undefined behavior. + * + * Example code: + * @code{.c} + * #include + * #include + * #include + * #define XXH_STATIC_LINKING_ONLY // expose unstable API + * #include "xxhash.h" + * // Hashes argv[2] using the entropy from argv[1]. + * int main(int argc, char* argv[]) + * { + * char secret[XXH3_SECRET_SIZE_MIN]; + * if (argv != 3) { return 1; } + * XXH3_generateSecret(secret, sizeof(secret), argv[1], strlen(argv[1])); + * XXH64_hash_t h = XXH3_64bits_withSecret( + * argv[2], strlen(argv[2]), + * secret, sizeof(secret) + * ); + * printf("%016llx\n", (unsigned long long) h); + * } + * @endcode + */ +XXH_PUBLIC_API XXH_errorcode XXH3_generateSecret(XXH_NOESCAPE void* secretBuffer, size_t secretSize, XXH_NOESCAPE const void* customSeed, size_t customSeedSize); + +/*! + * @brief Generate the same secret as the _withSeed() variants. + * + * @param secretBuffer A writable buffer of @ref XXH3_SECRET_DEFAULT_SIZE bytes + * @param seed The 64-bit seed to alter the hash result predictably. + * + * The generated secret can be used in combination with + *`*_withSecret()` and `_withSecretandSeed()` variants. + * + * Example C++ `std::string` hash class: + * @code{.cpp} + * #include + * #define XXH_STATIC_LINKING_ONLY // expose unstable API + * #include "xxhash.h" + * // Slow, seeds each time + * class HashSlow { + * XXH64_hash_t seed; + * public: + * HashSlow(XXH64_hash_t s) : seed{s} {} + * size_t operator()(const std::string& x) const { + * return size_t{XXH3_64bits_withSeed(x.c_str(), x.length(), seed)}; + * } + * }; + * // Fast, caches the seeded secret for future uses. + * class HashFast { + * unsigned char secret[XXH3_SECRET_DEFAULT_SIZE]; + * public: + * HashFast(XXH64_hash_t s) { + * XXH3_generateSecret_fromSeed(secret, seed); + * } + * size_t operator()(const std::string& x) const { + * return size_t{ + * XXH3_64bits_withSecret(x.c_str(), x.length(), secret, sizeof(secret)) + * }; + * } + * }; + * @endcode + */ +XXH_PUBLIC_API void XXH3_generateSecret_fromSeed(XXH_NOESCAPE void* secretBuffer, XXH64_hash_t seed); + +/*! + * @brief Maximum size of "short" key in bytes. + */ +#define XXH3_MIDSIZE_MAX 240 + +/*! + * @brief Calculates 64/128-bit seeded variant of XXH3 hash of @p data. + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * These variants generate hash values using either: + * - @p seed for "short" keys (< @ref XXH3_MIDSIZE_MAX = 240 bytes) + * - @p secret for "large" keys (>= @ref XXH3_MIDSIZE_MAX). + * + * This generally benefits speed, compared to `_withSeed()` or `_withSecret()`. + * `_withSeed()` has to generate the secret on the fly for "large" keys. + * It's fast, but can be perceptible for "not so large" keys (< 1 KB). + * `_withSecret()` has to generate the masks on the fly for "small" keys, + * which requires more instructions than _withSeed() variants. + * Therefore, _withSecretandSeed variant combines the best of both worlds. + * + * When @p secret has been generated by XXH3_generateSecret_fromSeed(), + * this variant produces *exactly* the same results as `_withSeed()` variant, + * hence offering only a pure speed benefit on "large" input, + * by skipping the need to regenerate the secret for every large input. + * + * Another usage scenario is to hash the secret to a 64-bit hash value, + * for example with XXH3_64bits(), which then becomes the seed, + * and then employ both the seed and the secret in _withSecretandSeed(). + * On top of speed, an added benefit is that each bit in the secret + * has a 50% chance to swap each bit in the output, via its impact to the seed. + * + * This is not guaranteed when using the secret directly in "small data" scenarios, + * because only portions of the secret are employed for small data. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t +XXH3_64bits_withSecretandSeed(XXH_NOESCAPE const void* data, size_t len, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed); + +/*! + * @brief Calculates 128-bit seeded variant of XXH3 hash of @p data. + * + * @param data The memory segment to be hashed, at least @p len bytes in size. + * @param length The length of @p data, in bytes. + * @param secret The secret used to alter hash result predictably. + * @param secretSize The length of @p secret, in bytes (must be >= XXH3_SECRET_SIZE_MIN) + * @param seed64 The 64-bit seed to alter the hash result predictably. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @see XXH3_64bits_withSecretandSeed(): contract is the same. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t +XXH3_128bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t length, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed64); + +#ifndef XXH_NO_STREAM +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * @param seed64 The 64-bit seed to alter the hash result predictably. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @see XXH3_64bits_withSecretandSeed(). Contract is identical. + */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed64); + +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * @param seed64 The 64-bit seed to alter the hash result predictably. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @see XXH3_64bits_withSecretandSeed(). Contract is identical. + * + * Note: there was a bug in an earlier version of this function (<= v0.8.2) + * that would make it generate an incorrect hash value + * when @p seed == 0 and @p length < XXH3_MIDSIZE_MAX + * and @p secret is different from XXH3_generateSecret_fromSeed(). + * As stated in the contract, the correct hash result must be + * the same as XXH3_128bits_withSeed() when @p length <= XXH3_MIDSIZE_MAX. + * Results generated by this older version are wrong, hence not comparable. + */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed64); + +#endif /* !XXH_NO_STREAM */ + +#endif /* !XXH_NO_XXH3 */ +#endif /* XXH_NO_LONG_LONG */ +#if defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API) +# define XXH_IMPLEMENTATION +#endif + +#endif /* defined(XXH_STATIC_LINKING_ONLY) && !defined(XXHASH_H_STATIC_13879238742) */ + + +/* ======================================================================== */ +/* ======================================================================== */ +/* ======================================================================== */ + + +/*-********************************************************************** + * xxHash implementation + *-********************************************************************** + * xxHash's implementation used to be hosted inside xxhash.c. + * + * However, inlining requires implementation to be visible to the compiler, + * hence be included alongside the header. + * Previously, implementation was hosted inside xxhash.c, + * which was then #included when inlining was activated. + * This construction created issues with a few build and install systems, + * as it required xxhash.c to be stored in /include directory. + * + * xxHash implementation is now directly integrated within xxhash.h. + * As a consequence, xxhash.c is no longer needed in /include. + * + * xxhash.c is still available and is still useful. + * In a "normal" setup, when xxhash is not inlined, + * xxhash.h only exposes the prototypes and public symbols, + * while xxhash.c can be built into an object file xxhash.o + * which can then be linked into the final binary. + ************************************************************************/ + +#if ( defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API) \ + || defined(XXH_IMPLEMENTATION) ) && !defined(XXH_IMPLEM_13a8737387) +# define XXH_IMPLEM_13a8737387 + +/* ************************************* +* Tuning parameters +***************************************/ + +/*! + * @defgroup tuning Tuning parameters + * @{ + * + * Various macros to control xxHash's behavior. + */ +#ifdef XXH_DOXYGEN +/*! + * @brief Define this to disable 64-bit code. + * + * Useful if only using the @ref XXH32_family and you have a strict C90 compiler. + */ +# define XXH_NO_LONG_LONG +# undef XXH_NO_LONG_LONG /* don't actually */ +/*! + * @brief Controls how unaligned memory is accessed. + * + * By default, access to unaligned memory is controlled by `memcpy()`, which is + * safe and portable. + * + * Unfortunately, on some target/compiler combinations, the generated assembly + * is sub-optimal. + * + * The below switch allow selection of a different access method + * in the search for improved performance. + * + * @par Possible options: + * + * - `XXH_FORCE_MEMORY_ACCESS=0` (default): `memcpy` + * @par + * Use `memcpy()`. Safe and portable. Note that most modern compilers will + * eliminate the function call and treat it as an unaligned access. + * + * - `XXH_FORCE_MEMORY_ACCESS=1`: `__attribute__((aligned(1)))` + * @par + * Depends on compiler extensions and is therefore not portable. + * This method is safe _if_ your compiler supports it, + * and *generally* as fast or faster than `memcpy`. + * + * - `XXH_FORCE_MEMORY_ACCESS=2`: Direct cast + * @par + * Casts directly and dereferences. This method doesn't depend on the + * compiler, but it violates the C standard as it directly dereferences an + * unaligned pointer. It can generate buggy code on targets which do not + * support unaligned memory accesses, but in some circumstances, it's the + * only known way to get the most performance. + * + * - `XXH_FORCE_MEMORY_ACCESS=3`: Byteshift + * @par + * Also portable. This can generate the best code on old compilers which don't + * inline small `memcpy()` calls, and it might also be faster on big-endian + * systems which lack a native byteswap instruction. However, some compilers + * will emit literal byteshifts even if the target supports unaligned access. + * + * + * @warning + * Methods 1 and 2 rely on implementation-defined behavior. Use these with + * care, as what works on one compiler/platform/optimization level may cause + * another to read garbage data or even crash. + * + * See https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html for details. + * + * Prefer these methods in priority order (0 > 3 > 1 > 2) + */ +# define XXH_FORCE_MEMORY_ACCESS 0 + +/*! + * @def XXH_SIZE_OPT + * @brief Controls how much xxHash optimizes for size. + * + * xxHash, when compiled, tends to result in a rather large binary size. This + * is mostly due to heavy usage to forced inlining and constant folding of the + * @ref XXH3_family to increase performance. + * + * However, some developers prefer size over speed. This option can + * significantly reduce the size of the generated code. When using the `-Os` + * or `-Oz` options on GCC or Clang, this is defined to 1 by default, + * otherwise it is defined to 0. + * + * Most of these size optimizations can be controlled manually. + * + * This is a number from 0-2. + * - `XXH_SIZE_OPT` == 0: Default. xxHash makes no size optimizations. Speed + * comes first. + * - `XXH_SIZE_OPT` == 1: Default for `-Os` and `-Oz`. xxHash is more + * conservative and disables hacks that increase code size. It implies the + * options @ref XXH_NO_INLINE_HINTS == 1, @ref XXH_FORCE_ALIGN_CHECK == 0, + * and @ref XXH3_NEON_LANES == 8 if they are not already defined. + * - `XXH_SIZE_OPT` == 2: xxHash tries to make itself as small as possible. + * Performance may cry. For example, the single shot functions just use the + * streaming API. + */ +# define XXH_SIZE_OPT 0 + +/*! + * @def XXH_FORCE_ALIGN_CHECK + * @brief If defined to non-zero, adds a special path for aligned inputs (XXH32() + * and XXH64() only). + * + * This is an important performance trick for architectures without decent + * unaligned memory access performance. + * + * It checks for input alignment, and when conditions are met, uses a "fast + * path" employing direct 32-bit/64-bit reads, resulting in _dramatically + * faster_ read speed. + * + * The check costs one initial branch per hash, which is generally negligible, + * but not zero. + * + * Moreover, it's not useful to generate an additional code path if memory + * access uses the same instruction for both aligned and unaligned + * addresses (e.g. x86 and aarch64). + * + * In these cases, the alignment check can be removed by setting this macro to 0. + * Then the code will always use unaligned memory access. + * Align check is automatically disabled on x86, x64, ARM64, and some ARM chips + * which are platforms known to offer good unaligned memory accesses performance. + * + * It is also disabled by default when @ref XXH_SIZE_OPT >= 1. + * + * This option does not affect XXH3 (only XXH32 and XXH64). + */ +# define XXH_FORCE_ALIGN_CHECK 0 + +/*! + * @def XXH_NO_INLINE_HINTS + * @brief When non-zero, sets all functions to `static`. + * + * By default, xxHash tries to force the compiler to inline almost all internal + * functions. + * + * This can usually improve performance due to reduced jumping and improved + * constant folding, but significantly increases the size of the binary which + * might not be favorable. + * + * Additionally, sometimes the forced inlining can be detrimental to performance, + * depending on the architecture. + * + * XXH_NO_INLINE_HINTS marks all internal functions as static, giving the + * compiler full control on whether to inline or not. + * + * When not optimizing (-O0), using `-fno-inline` with GCC or Clang, or if + * @ref XXH_SIZE_OPT >= 1, this will automatically be defined. + */ +# define XXH_NO_INLINE_HINTS 0 + +/*! + * @def XXH3_INLINE_SECRET + * @brief Determines whether to inline the XXH3 withSecret code. + * + * When the secret size is known, the compiler can improve the performance + * of XXH3_64bits_withSecret() and XXH3_128bits_withSecret(). + * + * However, if the secret size is not known, it doesn't have any benefit. This + * happens when xxHash is compiled into a global symbol. Therefore, if + * @ref XXH_INLINE_ALL is *not* defined, this will be defined to 0. + * + * Additionally, this defaults to 0 on GCC 12+, which has an issue with function pointers + * that are *sometimes* force inline on -Og, and it is impossible to automatically + * detect this optimization level. + */ +# define XXH3_INLINE_SECRET 0 + +/*! + * @def XXH32_ENDJMP + * @brief Whether to use a jump for `XXH32_finalize`. + * + * For performance, `XXH32_finalize` uses multiple branches in the finalizer. + * This is generally preferable for performance, + * but depending on exact architecture, a jmp may be preferable. + * + * This setting is only possibly making a difference for very small inputs. + */ +# define XXH32_ENDJMP 0 + +/*! + * @internal + * @brief Redefines old internal names. + * + * For compatibility with code that uses xxHash's internals before the names + * were changed to improve namespacing. There is no other reason to use this. + */ +# define XXH_OLD_NAMES +# undef XXH_OLD_NAMES /* don't actually use, it is ugly. */ + +/*! + * @def XXH_NO_STREAM + * @brief Disables the streaming API. + * + * When xxHash is not inlined and the streaming functions are not used, disabling + * the streaming functions can improve code size significantly, especially with + * the @ref XXH3_family which tends to make constant folded copies of itself. + */ +# define XXH_NO_STREAM +# undef XXH_NO_STREAM /* don't actually */ +#endif /* XXH_DOXYGEN */ +/*! + * @} + */ + +#ifndef XXH_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ + /* prefer __packed__ structures (method 1) for GCC + * < ARMv7 with unaligned access (e.g. Raspbian armhf) still uses byte shifting, so we use memcpy + * which for some reason does unaligned loads. */ +# if defined(__GNUC__) && !(defined(__ARM_ARCH) && __ARM_ARCH < 7 && defined(__ARM_FEATURE_UNALIGNED)) +# define XXH_FORCE_MEMORY_ACCESS 1 +# endif +#endif + +#ifndef XXH_SIZE_OPT + /* default to 1 for -Os or -Oz */ +# if (defined(__GNUC__) || defined(__clang__)) && defined(__OPTIMIZE_SIZE__) +# define XXH_SIZE_OPT 1 +# else +# define XXH_SIZE_OPT 0 +# endif +#endif + +#ifndef XXH_FORCE_ALIGN_CHECK /* can be defined externally */ + /* don't check on sizeopt, x86, aarch64, or arm when unaligned access is available */ +# if XXH_SIZE_OPT >= 1 || \ + defined(__i386) || defined(__x86_64__) || defined(__aarch64__) || defined(__ARM_FEATURE_UNALIGNED) \ + || defined(_M_IX86) || defined(_M_X64) || defined(_M_ARM64) || defined(_M_ARM) /* visual */ +# define XXH_FORCE_ALIGN_CHECK 0 +# else +# define XXH_FORCE_ALIGN_CHECK 1 +# endif +#endif + +#ifndef XXH_NO_INLINE_HINTS +# if XXH_SIZE_OPT >= 1 || defined(__NO_INLINE__) /* -O0, -fno-inline */ +# define XXH_NO_INLINE_HINTS 1 +# else +# define XXH_NO_INLINE_HINTS 0 +# endif +#endif + +#ifndef XXH3_INLINE_SECRET +# if (defined(__GNUC__) && !defined(__clang__) && __GNUC__ >= 12) \ + || !defined(XXH_INLINE_ALL) +# define XXH3_INLINE_SECRET 0 +# else +# define XXH3_INLINE_SECRET 1 +# endif +#endif + +#ifndef XXH32_ENDJMP +/* generally preferable for performance */ +# define XXH32_ENDJMP 0 +#endif + +/*! + * @defgroup impl Implementation + * @{ + */ + + +/* ************************************* +* Includes & Memory related functions +***************************************/ +#if defined(XXH_NO_STREAM) +/* nothing */ +#elif defined(XXH_NO_STDLIB) + +/* When requesting to disable any mention of stdlib, + * the library loses the ability to invoked malloc / free. + * In practice, it means that functions like `XXH*_createState()` + * will always fail, and return NULL. + * This flag is useful in situations where + * xxhash.h is integrated into some kernel, embedded or limited environment + * without access to dynamic allocation. + */ + +static XXH_CONSTF void* XXH_malloc(size_t s) { (void)s; return NULL; } +static void XXH_free(void* p) { (void)p; } + +#else + +/* + * Modify the local functions below should you wish to use + * different memory routines for malloc() and free() + */ +#include + +/*! + * @internal + * @brief Modify this function to use a different routine than malloc(). + */ +static XXH_MALLOCF void* XXH_malloc(size_t s) { return malloc(s); } + +/*! + * @internal + * @brief Modify this function to use a different routine than free(). + */ +static void XXH_free(void* p) { free(p); } + +#endif /* XXH_NO_STDLIB */ + +#include + +/*! + * @internal + * @brief Modify this function to use a different routine than memcpy(). + */ +static void* XXH_memcpy(void* dest, const void* src, size_t size) +{ + return memcpy(dest,src,size); +} + +#include /* ULLONG_MAX */ + + +/* ************************************* +* Compiler Specific Options +***************************************/ +#ifdef _MSC_VER /* Visual Studio warning fix */ +# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ +#endif + +#if XXH_NO_INLINE_HINTS /* disable inlining hints */ +# if defined(__GNUC__) || defined(__clang__) +# define XXH_FORCE_INLINE static __attribute__((__unused__)) +# else +# define XXH_FORCE_INLINE static +# endif +# define XXH_NO_INLINE static +/* enable inlining hints */ +#elif defined(__GNUC__) || defined(__clang__) +# define XXH_FORCE_INLINE static __inline__ __attribute__((__always_inline__, __unused__)) +# define XXH_NO_INLINE static __attribute__((__noinline__)) +#elif defined(_MSC_VER) /* Visual Studio */ +# define XXH_FORCE_INLINE static __forceinline +# define XXH_NO_INLINE static __declspec(noinline) +#elif defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L)) /* C99 */ +# define XXH_FORCE_INLINE static inline +# define XXH_NO_INLINE static +#else +# define XXH_FORCE_INLINE static +# define XXH_NO_INLINE static +#endif + +#if XXH3_INLINE_SECRET +# define XXH3_WITH_SECRET_INLINE XXH_FORCE_INLINE +#else +# define XXH3_WITH_SECRET_INLINE XXH_NO_INLINE +#endif + + +/* ************************************* +* Debug +***************************************/ +/*! + * @ingroup tuning + * @def XXH_DEBUGLEVEL + * @brief Sets the debugging level. + * + * XXH_DEBUGLEVEL is expected to be defined externally, typically via the + * compiler's command line options. The value must be a number. + */ +#ifndef XXH_DEBUGLEVEL +# ifdef DEBUGLEVEL /* backwards compat */ +# define XXH_DEBUGLEVEL DEBUGLEVEL +# else +# define XXH_DEBUGLEVEL 0 +# endif +#endif + +#if (XXH_DEBUGLEVEL>=1) +# include /* note: can still be disabled with NDEBUG */ +# define XXH_ASSERT(c) assert(c) +#else +# if defined(__INTEL_COMPILER) +# define XXH_ASSERT(c) XXH_ASSUME((unsigned char) (c)) +# else +# define XXH_ASSERT(c) XXH_ASSUME(c) +# endif +#endif + +/* note: use after variable declarations */ +#ifndef XXH_STATIC_ASSERT +# if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* C11 */ +# define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { _Static_assert((c),m); } while(0) +# elif defined(__cplusplus) && (__cplusplus >= 201103L) /* C++11 */ +# define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { static_assert((c),m); } while(0) +# else +# define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { struct xxh_sa { char x[(c) ? 1 : -1]; }; } while(0) +# endif +# define XXH_STATIC_ASSERT(c) XXH_STATIC_ASSERT_WITH_MESSAGE((c),#c) +#endif + +/*! + * @internal + * @def XXH_COMPILER_GUARD(var) + * @brief Used to prevent unwanted optimizations for @p var. + * + * It uses an empty GCC inline assembly statement with a register constraint + * which forces @p var into a general purpose register (eg eax, ebx, ecx + * on x86) and marks it as modified. + * + * This is used in a few places to avoid unwanted autovectorization (e.g. + * XXH32_round()). All vectorization we want is explicit via intrinsics, + * and _usually_ isn't wanted elsewhere. + * + * We also use it to prevent unwanted constant folding for AArch64 in + * XXH3_initCustomSecret_scalar(). + */ +#if defined(__GNUC__) || defined(__clang__) +# define XXH_COMPILER_GUARD(var) __asm__("" : "+r" (var)) +#else +# define XXH_COMPILER_GUARD(var) ((void)0) +#endif + +/* Specifically for NEON vectors which use the "w" constraint, on + * Clang. */ +#if defined(__clang__) && defined(__ARM_ARCH) && !defined(__wasm__) +# define XXH_COMPILER_GUARD_CLANG_NEON(var) __asm__("" : "+w" (var)) +#else +# define XXH_COMPILER_GUARD_CLANG_NEON(var) ((void)0) +#endif + +/* ************************************* +* Basic Types +***************************************/ +#if !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) +# ifdef _AIX +# include +# else +# include +# endif + typedef uint8_t xxh_u8; +#else + typedef unsigned char xxh_u8; +#endif +typedef XXH32_hash_t xxh_u32; + +#ifdef XXH_OLD_NAMES +# warning "XXH_OLD_NAMES is planned to be removed starting v0.9. If the program depends on it, consider moving away from it by employing newer type names directly" +# define BYTE xxh_u8 +# define U8 xxh_u8 +# define U32 xxh_u32 +#endif + +/* *** Memory access *** */ + +/*! + * @internal + * @fn xxh_u32 XXH_read32(const void* ptr) + * @brief Reads an unaligned 32-bit integer from @p ptr in native endianness. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * + * @param ptr The pointer to read from. + * @return The 32-bit native endian integer from the bytes at @p ptr. + */ + +/*! + * @internal + * @fn xxh_u32 XXH_readLE32(const void* ptr) + * @brief Reads an unaligned 32-bit little endian integer from @p ptr. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * + * @param ptr The pointer to read from. + * @return The 32-bit little endian integer from the bytes at @p ptr. + */ + +/*! + * @internal + * @fn xxh_u32 XXH_readBE32(const void* ptr) + * @brief Reads an unaligned 32-bit big endian integer from @p ptr. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * + * @param ptr The pointer to read from. + * @return The 32-bit big endian integer from the bytes at @p ptr. + */ + +/*! + * @internal + * @fn xxh_u32 XXH_readLE32_align(const void* ptr, XXH_alignment align) + * @brief Like @ref XXH_readLE32(), but has an option for aligned reads. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * Note that when @ref XXH_FORCE_ALIGN_CHECK == 0, the @p align parameter is + * always @ref XXH_alignment::XXH_unaligned. + * + * @param ptr The pointer to read from. + * @param align Whether @p ptr is aligned. + * @pre + * If @p align == @ref XXH_alignment::XXH_aligned, @p ptr must be 4 byte + * aligned. + * @return The 32-bit little endian integer from the bytes at @p ptr. + */ + +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) +/* + * Manual byteshift. Best for old compilers which don't inline memcpy. + * We actually directly use XXH_readLE32 and XXH_readBE32. + */ +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2)) + +/* + * Force direct memory access. Only works on CPU which support unaligned memory + * access in hardware. + */ +static xxh_u32 XXH_read32(const void* memPtr) { return *(const xxh_u32*) memPtr; } + +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1)) + +/* + * __attribute__((aligned(1))) is supported by gcc and clang. Originally the + * documentation claimed that it only increased the alignment, but actually it + * can decrease it on gcc, clang, and icc: + * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69502, + * https://gcc.godbolt.org/z/xYez1j67Y. + */ +#ifdef XXH_OLD_NAMES +typedef union { xxh_u32 u32; } __attribute__((__packed__)) unalign; +#endif +static xxh_u32 XXH_read32(const void* ptr) +{ + typedef __attribute__((__aligned__(1))) xxh_u32 xxh_unalign32; + return *((const xxh_unalign32*)ptr); +} + +#else + +/* + * Portable and safe solution. Generally efficient. + * see: https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html + */ +static xxh_u32 XXH_read32(const void* memPtr) +{ + xxh_u32 val; + XXH_memcpy(&val, memPtr, sizeof(val)); + return val; +} + +#endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */ + + +/* *** Endianness *** */ + +/*! + * @ingroup tuning + * @def XXH_CPU_LITTLE_ENDIAN + * @brief Whether the target is little endian. + * + * Defined to 1 if the target is little endian, or 0 if it is big endian. + * It can be defined externally, for example on the compiler command line. + * + * If it is not defined, + * a runtime check (which is usually constant folded) is used instead. + * + * @note + * This is not necessarily defined to an integer constant. + * + * @see XXH_isLittleEndian() for the runtime check. + */ +#ifndef XXH_CPU_LITTLE_ENDIAN +/* + * Try to detect endianness automatically, to avoid the nonstandard behavior + * in `XXH_isLittleEndian()` + */ +# if defined(_WIN32) /* Windows is always little endian */ \ + || defined(__LITTLE_ENDIAN__) \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) +# define XXH_CPU_LITTLE_ENDIAN 1 +# elif defined(__BIG_ENDIAN__) \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) +# define XXH_CPU_LITTLE_ENDIAN 0 +# else +/*! + * @internal + * @brief Runtime check for @ref XXH_CPU_LITTLE_ENDIAN. + * + * Most compilers will constant fold this. + */ +static int XXH_isLittleEndian(void) +{ + /* + * Portable and well-defined behavior. + * Don't use static: it is detrimental to performance. + */ + const union { xxh_u32 u; xxh_u8 c[4]; } one = { 1 }; + return one.c[0]; +} +# define XXH_CPU_LITTLE_ENDIAN XXH_isLittleEndian() +# endif +#endif + + + + +/* **************************************** +* Compiler-specific Functions and Macros +******************************************/ +#define XXH_GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) + +#ifdef __has_builtin +# define XXH_HAS_BUILTIN(x) __has_builtin(x) +#else +# define XXH_HAS_BUILTIN(x) 0 +#endif + + + +/* + * C23 and future versions have standard "unreachable()". + * Once it has been implemented reliably we can add it as an + * additional case: + * + * ``` + * #if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= XXH_C23_VN) + * # include + * # ifdef unreachable + * # define XXH_UNREACHABLE() unreachable() + * # endif + * #endif + * ``` + * + * Note C++23 also has std::unreachable() which can be detected + * as follows: + * ``` + * #if defined(__cpp_lib_unreachable) && (__cpp_lib_unreachable >= 202202L) + * # include + * # define XXH_UNREACHABLE() std::unreachable() + * #endif + * ``` + * NB: `__cpp_lib_unreachable` is defined in the `` header. + * We don't use that as including `` in `extern "C"` blocks + * doesn't work on GCC12 + */ + +#if XXH_HAS_BUILTIN(__builtin_unreachable) +# define XXH_UNREACHABLE() __builtin_unreachable() + +#elif defined(_MSC_VER) +# define XXH_UNREACHABLE() __assume(0) + +#else +# define XXH_UNREACHABLE() +#endif + +#if XXH_HAS_BUILTIN(__builtin_assume) +# define XXH_ASSUME(c) __builtin_assume(c) +#else +# define XXH_ASSUME(c) if (!(c)) { XXH_UNREACHABLE(); } +#endif + +/*! + * @internal + * @def XXH_rotl32(x,r) + * @brief 32-bit rotate left. + * + * @param x The 32-bit integer to be rotated. + * @param r The number of bits to rotate. + * @pre + * @p r > 0 && @p r < 32 + * @note + * @p x and @p r may be evaluated multiple times. + * @return The rotated result. + */ +#if !defined(NO_CLANG_BUILTIN) && XXH_HAS_BUILTIN(__builtin_rotateleft32) \ + && XXH_HAS_BUILTIN(__builtin_rotateleft64) +# define XXH_rotl32 __builtin_rotateleft32 +# define XXH_rotl64 __builtin_rotateleft64 +/* Note: although _rotl exists for minGW (GCC under windows), performance seems poor */ +#elif defined(_MSC_VER) +# define XXH_rotl32(x,r) _rotl(x,r) +# define XXH_rotl64(x,r) _rotl64(x,r) +#else +# define XXH_rotl32(x,r) (((x) << (r)) | ((x) >> (32 - (r)))) +# define XXH_rotl64(x,r) (((x) << (r)) | ((x) >> (64 - (r)))) +#endif + +/*! + * @internal + * @fn xxh_u32 XXH_swap32(xxh_u32 x) + * @brief A 32-bit byteswap. + * + * @param x The 32-bit integer to byteswap. + * @return @p x, byteswapped. + */ +#if defined(_MSC_VER) /* Visual Studio */ +# define XXH_swap32 _byteswap_ulong +#elif XXH_GCC_VERSION >= 403 +# define XXH_swap32 __builtin_bswap32 +#else +static xxh_u32 XXH_swap32 (xxh_u32 x) +{ + return ((x << 24) & 0xff000000 ) | + ((x << 8) & 0x00ff0000 ) | + ((x >> 8) & 0x0000ff00 ) | + ((x >> 24) & 0x000000ff ); +} +#endif + + +/* *************************** +* Memory reads +*****************************/ + +/*! + * @internal + * @brief Enum to indicate whether a pointer is aligned. + */ +typedef enum { + XXH_aligned, /*!< Aligned */ + XXH_unaligned /*!< Possibly unaligned */ +} XXH_alignment; + +/* + * XXH_FORCE_MEMORY_ACCESS==3 is an endian-independent byteshift load. + * + * This is ideal for older compilers which don't inline memcpy. + */ +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) + +XXH_FORCE_INLINE xxh_u32 XXH_readLE32(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[0] + | ((xxh_u32)bytePtr[1] << 8) + | ((xxh_u32)bytePtr[2] << 16) + | ((xxh_u32)bytePtr[3] << 24); +} + +XXH_FORCE_INLINE xxh_u32 XXH_readBE32(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[3] + | ((xxh_u32)bytePtr[2] << 8) + | ((xxh_u32)bytePtr[1] << 16) + | ((xxh_u32)bytePtr[0] << 24); +} + +#else +XXH_FORCE_INLINE xxh_u32 XXH_readLE32(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_read32(ptr) : XXH_swap32(XXH_read32(ptr)); +} + +static xxh_u32 XXH_readBE32(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_swap32(XXH_read32(ptr)) : XXH_read32(ptr); +} +#endif + +XXH_FORCE_INLINE xxh_u32 +XXH_readLE32_align(const void* ptr, XXH_alignment align) +{ + if (align==XXH_unaligned) { + return XXH_readLE32(ptr); + } else { + return XXH_CPU_LITTLE_ENDIAN ? *(const xxh_u32*)ptr : XXH_swap32(*(const xxh_u32*)ptr); + } +} + + +/* ************************************* +* Misc +***************************************/ +/*! @ingroup public */ +XXH_PUBLIC_API unsigned XXH_versionNumber (void) { return XXH_VERSION_NUMBER; } + + +/* ******************************************************************* +* 32-bit hash functions +*********************************************************************/ +/*! + * @} + * @defgroup XXH32_impl XXH32 implementation + * @ingroup impl + * + * Details on the XXH32 implementation. + * @{ + */ + /* #define instead of static const, to be used as initializers */ +#define XXH_PRIME32_1 0x9E3779B1U /*!< 0b10011110001101110111100110110001 */ +#define XXH_PRIME32_2 0x85EBCA77U /*!< 0b10000101111010111100101001110111 */ +#define XXH_PRIME32_3 0xC2B2AE3DU /*!< 0b11000010101100101010111000111101 */ +#define XXH_PRIME32_4 0x27D4EB2FU /*!< 0b00100111110101001110101100101111 */ +#define XXH_PRIME32_5 0x165667B1U /*!< 0b00010110010101100110011110110001 */ + +#ifdef XXH_OLD_NAMES +# define PRIME32_1 XXH_PRIME32_1 +# define PRIME32_2 XXH_PRIME32_2 +# define PRIME32_3 XXH_PRIME32_3 +# define PRIME32_4 XXH_PRIME32_4 +# define PRIME32_5 XXH_PRIME32_5 +#endif + +/*! + * @internal + * @brief Normal stripe processing routine. + * + * This shuffles the bits so that any bit from @p input impacts several bits in + * @p acc. + * + * @param acc The accumulator lane. + * @param input The stripe of input to mix. + * @return The mixed accumulator lane. + */ +static xxh_u32 XXH32_round(xxh_u32 acc, xxh_u32 input) +{ + acc += input * XXH_PRIME32_2; + acc = XXH_rotl32(acc, 13); + acc *= XXH_PRIME32_1; +#if (defined(__SSE4_1__) || defined(__aarch64__) || defined(__wasm_simd128__)) && !defined(XXH_ENABLE_AUTOVECTORIZE) + /* + * UGLY HACK: + * A compiler fence is used to prevent GCC and Clang from + * autovectorizing the XXH32 loop (pragmas and attributes don't work for some + * reason) without globally disabling SSE4.1. + * + * The reason we want to avoid vectorization is because despite working on + * 4 integers at a time, there are multiple factors slowing XXH32 down on + * SSE4: + * - There's a ridiculous amount of lag from pmulld (10 cycles of latency on + * newer chips!) making it slightly slower to multiply four integers at + * once compared to four integers independently. Even when pmulld was + * fastest, Sandy/Ivy Bridge, it is still not worth it to go into SSE + * just to multiply unless doing a long operation. + * + * - Four instructions are required to rotate, + * movqda tmp, v // not required with VEX encoding + * pslld tmp, 13 // tmp <<= 13 + * psrld v, 19 // x >>= 19 + * por v, tmp // x |= tmp + * compared to one for scalar: + * roll v, 13 // reliably fast across the board + * shldl v, v, 13 // Sandy Bridge and later prefer this for some reason + * + * - Instruction level parallelism is actually more beneficial here because + * the SIMD actually serializes this operation: While v1 is rotating, v2 + * can load data, while v3 can multiply. SSE forces them to operate + * together. + * + * This is also enabled on AArch64, as Clang is *very aggressive* in vectorizing + * the loop. NEON is only faster on the A53, and with the newer cores, it is less + * than half the speed. + * + * Additionally, this is used on WASM SIMD128 because it JITs to the same + * SIMD instructions and has the same issue. + */ + XXH_COMPILER_GUARD(acc); +#endif + return acc; +} + +/*! + * @internal + * @brief Mixes all bits to finalize the hash. + * + * The final mix ensures that all input bits have a chance to impact any bit in + * the output digest, resulting in an unbiased distribution. + * + * @param hash The hash to avalanche. + * @return The avalanched hash. + */ +static xxh_u32 XXH32_avalanche(xxh_u32 hash) +{ + hash ^= hash >> 15; + hash *= XXH_PRIME32_2; + hash ^= hash >> 13; + hash *= XXH_PRIME32_3; + hash ^= hash >> 16; + return hash; +} + +#define XXH_get32bits(p) XXH_readLE32_align(p, align) + +/*! + * @internal + * @brief Processes the last 0-15 bytes of @p ptr. + * + * There may be up to 15 bytes remaining to consume from the input. + * This final stage will digest them to ensure that all input bytes are present + * in the final mix. + * + * @param hash The hash to finalize. + * @param ptr The pointer to the remaining input. + * @param len The remaining length, modulo 16. + * @param align Whether @p ptr is aligned. + * @return The finalized hash. + * @see XXH64_finalize(). + */ +static XXH_PUREF xxh_u32 +XXH32_finalize(xxh_u32 hash, const xxh_u8* ptr, size_t len, XXH_alignment align) +{ +#define XXH_PROCESS1 do { \ + hash += (*ptr++) * XXH_PRIME32_5; \ + hash = XXH_rotl32(hash, 11) * XXH_PRIME32_1; \ +} while (0) + +#define XXH_PROCESS4 do { \ + hash += XXH_get32bits(ptr) * XXH_PRIME32_3; \ + ptr += 4; \ + hash = XXH_rotl32(hash, 17) * XXH_PRIME32_4; \ +} while (0) + + if (ptr==NULL) XXH_ASSERT(len == 0); + + /* Compact rerolled version; generally faster */ + if (!XXH32_ENDJMP) { + len &= 15; + while (len >= 4) { + XXH_PROCESS4; + len -= 4; + } + while (len > 0) { + XXH_PROCESS1; + --len; + } + return XXH32_avalanche(hash); + } else { + switch(len&15) /* or switch(bEnd - p) */ { + case 12: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 8: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 4: XXH_PROCESS4; + return XXH32_avalanche(hash); + + case 13: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 9: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 5: XXH_PROCESS4; + XXH_PROCESS1; + return XXH32_avalanche(hash); + + case 14: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 10: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 6: XXH_PROCESS4; + XXH_PROCESS1; + XXH_PROCESS1; + return XXH32_avalanche(hash); + + case 15: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 11: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 7: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 3: XXH_PROCESS1; + XXH_FALLTHROUGH; /* fallthrough */ + case 2: XXH_PROCESS1; + XXH_FALLTHROUGH; /* fallthrough */ + case 1: XXH_PROCESS1; + XXH_FALLTHROUGH; /* fallthrough */ + case 0: return XXH32_avalanche(hash); + } + XXH_ASSERT(0); + return hash; /* reaching this point is deemed impossible */ + } +} + +#ifdef XXH_OLD_NAMES +# define PROCESS1 XXH_PROCESS1 +# define PROCESS4 XXH_PROCESS4 +#else +# undef XXH_PROCESS1 +# undef XXH_PROCESS4 +#endif + +/*! + * @internal + * @brief The implementation for @ref XXH32(). + * + * @param input , len , seed Directly passed from @ref XXH32(). + * @param align Whether @p input is aligned. + * @return The calculated hash. + */ +XXH_FORCE_INLINE XXH_PUREF xxh_u32 +XXH32_endian_align(const xxh_u8* input, size_t len, xxh_u32 seed, XXH_alignment align) +{ + xxh_u32 h32; + + if (input==NULL) XXH_ASSERT(len == 0); + + if (len>=16) { + const xxh_u8* const bEnd = input + len; + const xxh_u8* const limit = bEnd - 15; + xxh_u32 v1 = seed + XXH_PRIME32_1 + XXH_PRIME32_2; + xxh_u32 v2 = seed + XXH_PRIME32_2; + xxh_u32 v3 = seed + 0; + xxh_u32 v4 = seed - XXH_PRIME32_1; + + do { + v1 = XXH32_round(v1, XXH_get32bits(input)); input += 4; + v2 = XXH32_round(v2, XXH_get32bits(input)); input += 4; + v3 = XXH32_round(v3, XXH_get32bits(input)); input += 4; + v4 = XXH32_round(v4, XXH_get32bits(input)); input += 4; + } while (input < limit); + + h32 = XXH_rotl32(v1, 1) + XXH_rotl32(v2, 7) + + XXH_rotl32(v3, 12) + XXH_rotl32(v4, 18); + } else { + h32 = seed + XXH_PRIME32_5; + } + + h32 += (xxh_u32)len; + + return XXH32_finalize(h32, input, len&15, align); +} + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_hash_t XXH32 (const void* input, size_t len, XXH32_hash_t seed) +{ +#if !defined(XXH_NO_STREAM) && XXH_SIZE_OPT >= 2 + /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ + XXH32_state_t state; + XXH32_reset(&state, seed); + XXH32_update(&state, (const xxh_u8*)input, len); + return XXH32_digest(&state); +#else + if (XXH_FORCE_ALIGN_CHECK) { + if ((((size_t)input) & 3) == 0) { /* Input is 4-bytes aligned, leverage the speed benefit */ + return XXH32_endian_align((const xxh_u8*)input, len, seed, XXH_aligned); + } } + + return XXH32_endian_align((const xxh_u8*)input, len, seed, XXH_unaligned); +#endif +} + + + +/******* Hash streaming *******/ +#ifndef XXH_NO_STREAM +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_state_t* XXH32_createState(void) +{ + return (XXH32_state_t*)XXH_malloc(sizeof(XXH32_state_t)); +} +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr) +{ + XXH_free(statePtr); + return XXH_OK; +} + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dstState, const XXH32_state_t* srcState) +{ + XXH_memcpy(dstState, srcState, sizeof(*dstState)); +} + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH_errorcode XXH32_reset(XXH32_state_t* statePtr, XXH32_hash_t seed) +{ + XXH_ASSERT(statePtr != NULL); + memset(statePtr, 0, sizeof(*statePtr)); + statePtr->v[0] = seed + XXH_PRIME32_1 + XXH_PRIME32_2; + statePtr->v[1] = seed + XXH_PRIME32_2; + statePtr->v[2] = seed + 0; + statePtr->v[3] = seed - XXH_PRIME32_1; + return XXH_OK; +} + + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH_errorcode +XXH32_update(XXH32_state_t* state, const void* input, size_t len) +{ + if (input==NULL) { + XXH_ASSERT(len == 0); + return XXH_OK; + } + + { const xxh_u8* p = (const xxh_u8*)input; + const xxh_u8* const bEnd = p + len; + + state->total_len_32 += (XXH32_hash_t)len; + state->large_len |= (XXH32_hash_t)((len>=16) | (state->total_len_32>=16)); + + if (state->memsize + len < 16) { /* fill in tmp buffer */ + XXH_memcpy((xxh_u8*)(state->mem32) + state->memsize, input, len); + state->memsize += (XXH32_hash_t)len; + return XXH_OK; + } + + if (state->memsize) { /* some data left from previous update */ + XXH_memcpy((xxh_u8*)(state->mem32) + state->memsize, input, 16-state->memsize); + { const xxh_u32* p32 = state->mem32; + state->v[0] = XXH32_round(state->v[0], XXH_readLE32(p32)); p32++; + state->v[1] = XXH32_round(state->v[1], XXH_readLE32(p32)); p32++; + state->v[2] = XXH32_round(state->v[2], XXH_readLE32(p32)); p32++; + state->v[3] = XXH32_round(state->v[3], XXH_readLE32(p32)); + } + p += 16-state->memsize; + state->memsize = 0; + } + + if (p <= bEnd-16) { + const xxh_u8* const limit = bEnd - 16; + + do { + state->v[0] = XXH32_round(state->v[0], XXH_readLE32(p)); p+=4; + state->v[1] = XXH32_round(state->v[1], XXH_readLE32(p)); p+=4; + state->v[2] = XXH32_round(state->v[2], XXH_readLE32(p)); p+=4; + state->v[3] = XXH32_round(state->v[3], XXH_readLE32(p)); p+=4; + } while (p<=limit); + + } + + if (p < bEnd) { + XXH_memcpy(state->mem32, p, (size_t)(bEnd-p)); + state->memsize = (unsigned)(bEnd-p); + } + } + + return XXH_OK; +} + + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_hash_t XXH32_digest(const XXH32_state_t* state) +{ + xxh_u32 h32; + + if (state->large_len) { + h32 = XXH_rotl32(state->v[0], 1) + + XXH_rotl32(state->v[1], 7) + + XXH_rotl32(state->v[2], 12) + + XXH_rotl32(state->v[3], 18); + } else { + h32 = state->v[2] /* == seed */ + XXH_PRIME32_5; + } + + h32 += state->total_len_32; + + return XXH32_finalize(h32, (const xxh_u8*)state->mem32, state->memsize, XXH_aligned); +} +#endif /* !XXH_NO_STREAM */ + +/******* Canonical representation *******/ + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash) +{ + XXH_STATIC_ASSERT(sizeof(XXH32_canonical_t) == sizeof(XXH32_hash_t)); + if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap32(hash); + XXH_memcpy(dst, &hash, sizeof(*dst)); +} +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src) +{ + return XXH_readBE32(src); +} + + +#ifndef XXH_NO_LONG_LONG + +/* ******************************************************************* +* 64-bit hash functions +*********************************************************************/ +/*! + * @} + * @ingroup impl + * @{ + */ +/******* Memory access *******/ + +typedef XXH64_hash_t xxh_u64; + +#ifdef XXH_OLD_NAMES +# define U64 xxh_u64 +#endif + +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) +/* + * Manual byteshift. Best for old compilers which don't inline memcpy. + * We actually directly use XXH_readLE64 and XXH_readBE64. + */ +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2)) + +/* Force direct memory access. Only works on CPU which support unaligned memory access in hardware */ +static xxh_u64 XXH_read64(const void* memPtr) +{ + return *(const xxh_u64*) memPtr; +} + +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1)) + +/* + * __attribute__((aligned(1))) is supported by gcc and clang. Originally the + * documentation claimed that it only increased the alignment, but actually it + * can decrease it on gcc, clang, and icc: + * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69502, + * https://gcc.godbolt.org/z/xYez1j67Y. + */ +#ifdef XXH_OLD_NAMES +typedef union { xxh_u32 u32; xxh_u64 u64; } __attribute__((__packed__)) unalign64; +#endif +static xxh_u64 XXH_read64(const void* ptr) +{ + typedef __attribute__((__aligned__(1))) xxh_u64 xxh_unalign64; + return *((const xxh_unalign64*)ptr); +} + +#else + +/* + * Portable and safe solution. Generally efficient. + * see: https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html + */ +static xxh_u64 XXH_read64(const void* memPtr) +{ + xxh_u64 val; + XXH_memcpy(&val, memPtr, sizeof(val)); + return val; +} + +#endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */ + +#if defined(_MSC_VER) /* Visual Studio */ +# define XXH_swap64 _byteswap_uint64 +#elif XXH_GCC_VERSION >= 403 +# define XXH_swap64 __builtin_bswap64 +#else +static xxh_u64 XXH_swap64(xxh_u64 x) +{ + return ((x << 56) & 0xff00000000000000ULL) | + ((x << 40) & 0x00ff000000000000ULL) | + ((x << 24) & 0x0000ff0000000000ULL) | + ((x << 8) & 0x000000ff00000000ULL) | + ((x >> 8) & 0x00000000ff000000ULL) | + ((x >> 24) & 0x0000000000ff0000ULL) | + ((x >> 40) & 0x000000000000ff00ULL) | + ((x >> 56) & 0x00000000000000ffULL); +} +#endif + + +/* XXH_FORCE_MEMORY_ACCESS==3 is an endian-independent byteshift load. */ +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) + +XXH_FORCE_INLINE xxh_u64 XXH_readLE64(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[0] + | ((xxh_u64)bytePtr[1] << 8) + | ((xxh_u64)bytePtr[2] << 16) + | ((xxh_u64)bytePtr[3] << 24) + | ((xxh_u64)bytePtr[4] << 32) + | ((xxh_u64)bytePtr[5] << 40) + | ((xxh_u64)bytePtr[6] << 48) + | ((xxh_u64)bytePtr[7] << 56); +} + +XXH_FORCE_INLINE xxh_u64 XXH_readBE64(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[7] + | ((xxh_u64)bytePtr[6] << 8) + | ((xxh_u64)bytePtr[5] << 16) + | ((xxh_u64)bytePtr[4] << 24) + | ((xxh_u64)bytePtr[3] << 32) + | ((xxh_u64)bytePtr[2] << 40) + | ((xxh_u64)bytePtr[1] << 48) + | ((xxh_u64)bytePtr[0] << 56); +} + +#else +XXH_FORCE_INLINE xxh_u64 XXH_readLE64(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_read64(ptr) : XXH_swap64(XXH_read64(ptr)); +} + +static xxh_u64 XXH_readBE64(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_swap64(XXH_read64(ptr)) : XXH_read64(ptr); +} +#endif + +XXH_FORCE_INLINE xxh_u64 +XXH_readLE64_align(const void* ptr, XXH_alignment align) +{ + if (align==XXH_unaligned) + return XXH_readLE64(ptr); + else + return XXH_CPU_LITTLE_ENDIAN ? *(const xxh_u64*)ptr : XXH_swap64(*(const xxh_u64*)ptr); +} + + +/******* xxh64 *******/ +/*! + * @} + * @defgroup XXH64_impl XXH64 implementation + * @ingroup impl + * + * Details on the XXH64 implementation. + * @{ + */ +/* #define rather that static const, to be used as initializers */ +#define XXH_PRIME64_1 0x9E3779B185EBCA87ULL /*!< 0b1001111000110111011110011011000110000101111010111100101010000111 */ +#define XXH_PRIME64_2 0xC2B2AE3D27D4EB4FULL /*!< 0b1100001010110010101011100011110100100111110101001110101101001111 */ +#define XXH_PRIME64_3 0x165667B19E3779F9ULL /*!< 0b0001011001010110011001111011000110011110001101110111100111111001 */ +#define XXH_PRIME64_4 0x85EBCA77C2B2AE63ULL /*!< 0b1000010111101011110010100111011111000010101100101010111001100011 */ +#define XXH_PRIME64_5 0x27D4EB2F165667C5ULL /*!< 0b0010011111010100111010110010111100010110010101100110011111000101 */ + +#ifdef XXH_OLD_NAMES +# define PRIME64_1 XXH_PRIME64_1 +# define PRIME64_2 XXH_PRIME64_2 +# define PRIME64_3 XXH_PRIME64_3 +# define PRIME64_4 XXH_PRIME64_4 +# define PRIME64_5 XXH_PRIME64_5 +#endif + +/*! @copydoc XXH32_round */ +static xxh_u64 XXH64_round(xxh_u64 acc, xxh_u64 input) +{ + acc += input * XXH_PRIME64_2; + acc = XXH_rotl64(acc, 31); + acc *= XXH_PRIME64_1; +#if (defined(__AVX512F__)) && !defined(XXH_ENABLE_AUTOVECTORIZE) + /* + * DISABLE AUTOVECTORIZATION: + * A compiler fence is used to prevent GCC and Clang from + * autovectorizing the XXH64 loop (pragmas and attributes don't work for some + * reason) without globally disabling AVX512. + * + * Autovectorization of XXH64 tends to be detrimental, + * though the exact outcome may change depending on exact cpu and compiler version. + * For information, it has been reported as detrimental for Skylake-X, + * but possibly beneficial for Zen4. + * + * The default is to disable auto-vectorization, + * but you can select to enable it instead using `XXH_ENABLE_AUTOVECTORIZE` build variable. + */ + XXH_COMPILER_GUARD(acc); +#endif + return acc; +} + +static xxh_u64 XXH64_mergeRound(xxh_u64 acc, xxh_u64 val) +{ + val = XXH64_round(0, val); + acc ^= val; + acc = acc * XXH_PRIME64_1 + XXH_PRIME64_4; + return acc; +} + +/*! @copydoc XXH32_avalanche */ +static xxh_u64 XXH64_avalanche(xxh_u64 hash) +{ + hash ^= hash >> 33; + hash *= XXH_PRIME64_2; + hash ^= hash >> 29; + hash *= XXH_PRIME64_3; + hash ^= hash >> 32; + return hash; +} + + +#define XXH_get64bits(p) XXH_readLE64_align(p, align) + +/*! + * @internal + * @brief Processes the last 0-31 bytes of @p ptr. + * + * There may be up to 31 bytes remaining to consume from the input. + * This final stage will digest them to ensure that all input bytes are present + * in the final mix. + * + * @param hash The hash to finalize. + * @param ptr The pointer to the remaining input. + * @param len The remaining length, modulo 32. + * @param align Whether @p ptr is aligned. + * @return The finalized hash + * @see XXH32_finalize(). + */ +static XXH_PUREF xxh_u64 +XXH64_finalize(xxh_u64 hash, const xxh_u8* ptr, size_t len, XXH_alignment align) +{ + if (ptr==NULL) XXH_ASSERT(len == 0); + len &= 31; + while (len >= 8) { + xxh_u64 const k1 = XXH64_round(0, XXH_get64bits(ptr)); + ptr += 8; + hash ^= k1; + hash = XXH_rotl64(hash,27) * XXH_PRIME64_1 + XXH_PRIME64_4; + len -= 8; + } + if (len >= 4) { + hash ^= (xxh_u64)(XXH_get32bits(ptr)) * XXH_PRIME64_1; + ptr += 4; + hash = XXH_rotl64(hash, 23) * XXH_PRIME64_2 + XXH_PRIME64_3; + len -= 4; + } + while (len > 0) { + hash ^= (*ptr++) * XXH_PRIME64_5; + hash = XXH_rotl64(hash, 11) * XXH_PRIME64_1; + --len; + } + return XXH64_avalanche(hash); +} + +#ifdef XXH_OLD_NAMES +# define PROCESS1_64 XXH_PROCESS1_64 +# define PROCESS4_64 XXH_PROCESS4_64 +# define PROCESS8_64 XXH_PROCESS8_64 +#else +# undef XXH_PROCESS1_64 +# undef XXH_PROCESS4_64 +# undef XXH_PROCESS8_64 +#endif + +/*! + * @internal + * @brief The implementation for @ref XXH64(). + * + * @param input , len , seed Directly passed from @ref XXH64(). + * @param align Whether @p input is aligned. + * @return The calculated hash. + */ +XXH_FORCE_INLINE XXH_PUREF xxh_u64 +XXH64_endian_align(const xxh_u8* input, size_t len, xxh_u64 seed, XXH_alignment align) +{ + xxh_u64 h64; + if (input==NULL) XXH_ASSERT(len == 0); + + if (len>=32) { + const xxh_u8* const bEnd = input + len; + const xxh_u8* const limit = bEnd - 31; + xxh_u64 v1 = seed + XXH_PRIME64_1 + XXH_PRIME64_2; + xxh_u64 v2 = seed + XXH_PRIME64_2; + xxh_u64 v3 = seed + 0; + xxh_u64 v4 = seed - XXH_PRIME64_1; + + do { + v1 = XXH64_round(v1, XXH_get64bits(input)); input+=8; + v2 = XXH64_round(v2, XXH_get64bits(input)); input+=8; + v3 = XXH64_round(v3, XXH_get64bits(input)); input+=8; + v4 = XXH64_round(v4, XXH_get64bits(input)); input+=8; + } while (input= 2 + /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ + XXH64_state_t state; + XXH64_reset(&state, seed); + XXH64_update(&state, (const xxh_u8*)input, len); + return XXH64_digest(&state); +#else + if (XXH_FORCE_ALIGN_CHECK) { + if ((((size_t)input) & 7)==0) { /* Input is aligned, let's leverage the speed advantage */ + return XXH64_endian_align((const xxh_u8*)input, len, seed, XXH_aligned); + } } + + return XXH64_endian_align((const xxh_u8*)input, len, seed, XXH_unaligned); + +#endif +} + +/******* Hash Streaming *******/ +#ifndef XXH_NO_STREAM +/*! @ingroup XXH64_family*/ +XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void) +{ + return (XXH64_state_t*)XXH_malloc(sizeof(XXH64_state_t)); +} +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr) +{ + XXH_free(statePtr); + return XXH_OK; +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API void XXH64_copyState(XXH_NOESCAPE XXH64_state_t* dstState, const XXH64_state_t* srcState) +{ + XXH_memcpy(dstState, srcState, sizeof(*dstState)); +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH_errorcode XXH64_reset(XXH_NOESCAPE XXH64_state_t* statePtr, XXH64_hash_t seed) +{ + XXH_ASSERT(statePtr != NULL); + memset(statePtr, 0, sizeof(*statePtr)); + statePtr->v[0] = seed + XXH_PRIME64_1 + XXH_PRIME64_2; + statePtr->v[1] = seed + XXH_PRIME64_2; + statePtr->v[2] = seed + 0; + statePtr->v[3] = seed - XXH_PRIME64_1; + return XXH_OK; +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH_errorcode +XXH64_update (XXH_NOESCAPE XXH64_state_t* state, XXH_NOESCAPE const void* input, size_t len) +{ + if (input==NULL) { + XXH_ASSERT(len == 0); + return XXH_OK; + } + + { const xxh_u8* p = (const xxh_u8*)input; + const xxh_u8* const bEnd = p + len; + + state->total_len += len; + + if (state->memsize + len < 32) { /* fill in tmp buffer */ + XXH_memcpy(((xxh_u8*)state->mem64) + state->memsize, input, len); + state->memsize += (xxh_u32)len; + return XXH_OK; + } + + if (state->memsize) { /* tmp buffer is full */ + XXH_memcpy(((xxh_u8*)state->mem64) + state->memsize, input, 32-state->memsize); + state->v[0] = XXH64_round(state->v[0], XXH_readLE64(state->mem64+0)); + state->v[1] = XXH64_round(state->v[1], XXH_readLE64(state->mem64+1)); + state->v[2] = XXH64_round(state->v[2], XXH_readLE64(state->mem64+2)); + state->v[3] = XXH64_round(state->v[3], XXH_readLE64(state->mem64+3)); + p += 32 - state->memsize; + state->memsize = 0; + } + + if (p+32 <= bEnd) { + const xxh_u8* const limit = bEnd - 32; + + do { + state->v[0] = XXH64_round(state->v[0], XXH_readLE64(p)); p+=8; + state->v[1] = XXH64_round(state->v[1], XXH_readLE64(p)); p+=8; + state->v[2] = XXH64_round(state->v[2], XXH_readLE64(p)); p+=8; + state->v[3] = XXH64_round(state->v[3], XXH_readLE64(p)); p+=8; + } while (p<=limit); + + } + + if (p < bEnd) { + XXH_memcpy(state->mem64, p, (size_t)(bEnd-p)); + state->memsize = (unsigned)(bEnd-p); + } + } + + return XXH_OK; +} + + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH64_hash_t XXH64_digest(XXH_NOESCAPE const XXH64_state_t* state) +{ + xxh_u64 h64; + + if (state->total_len >= 32) { + h64 = XXH_rotl64(state->v[0], 1) + XXH_rotl64(state->v[1], 7) + XXH_rotl64(state->v[2], 12) + XXH_rotl64(state->v[3], 18); + h64 = XXH64_mergeRound(h64, state->v[0]); + h64 = XXH64_mergeRound(h64, state->v[1]); + h64 = XXH64_mergeRound(h64, state->v[2]); + h64 = XXH64_mergeRound(h64, state->v[3]); + } else { + h64 = state->v[2] /*seed*/ + XXH_PRIME64_5; + } + + h64 += (xxh_u64) state->total_len; + + return XXH64_finalize(h64, (const xxh_u8*)state->mem64, (size_t)state->total_len, XXH_aligned); +} +#endif /* !XXH_NO_STREAM */ + +/******* Canonical representation *******/ + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH_NOESCAPE XXH64_canonical_t* dst, XXH64_hash_t hash) +{ + XXH_STATIC_ASSERT(sizeof(XXH64_canonical_t) == sizeof(XXH64_hash_t)); + if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap64(hash); + XXH_memcpy(dst, &hash, sizeof(*dst)); +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(XXH_NOESCAPE const XXH64_canonical_t* src) +{ + return XXH_readBE64(src); +} + +#ifndef XXH_NO_XXH3 + +/* ********************************************************************* +* XXH3 +* New generation hash designed for speed on small keys and vectorization +************************************************************************ */ +/*! + * @} + * @defgroup XXH3_impl XXH3 implementation + * @ingroup impl + * @{ + */ + +/* === Compiler specifics === */ + +#if ((defined(sun) || defined(__sun)) && __cplusplus) /* Solaris includes __STDC_VERSION__ with C++. Tested with GCC 5.5 */ +# define XXH_RESTRICT /* disable */ +#elif defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* >= C99 */ +# define XXH_RESTRICT restrict +#elif (defined (__GNUC__) && ((__GNUC__ > 3) || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1))) \ + || (defined (__clang__)) \ + || (defined (_MSC_VER) && (_MSC_VER >= 1400)) \ + || (defined (__INTEL_COMPILER) && (__INTEL_COMPILER >= 1300)) +/* + * There are a LOT more compilers that recognize __restrict but this + * covers the major ones. + */ +# define XXH_RESTRICT __restrict +#else +# define XXH_RESTRICT /* disable */ +#endif + +#if (defined(__GNUC__) && (__GNUC__ >= 3)) \ + || (defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 800)) \ + || defined(__clang__) +# define XXH_likely(x) __builtin_expect(x, 1) +# define XXH_unlikely(x) __builtin_expect(x, 0) +#else +# define XXH_likely(x) (x) +# define XXH_unlikely(x) (x) +#endif + +#ifndef XXH_HAS_INCLUDE +# ifdef __has_include +/* + * Not defined as XXH_HAS_INCLUDE(x) (function-like) because + * this causes segfaults in Apple Clang 4.2 (on Mac OS X 10.7 Lion) + */ +# define XXH_HAS_INCLUDE __has_include +# else +# define XXH_HAS_INCLUDE(x) 0 +# endif +#endif + +#if defined(__GNUC__) || defined(__clang__) +# if defined(__ARM_FEATURE_SVE) +# include +# endif +# if defined(__ARM_NEON__) || defined(__ARM_NEON) \ + || (defined(_M_ARM) && _M_ARM >= 7) \ + || defined(_M_ARM64) || defined(_M_ARM64EC) \ + || (defined(__wasm_simd128__) && XXH_HAS_INCLUDE()) /* WASM SIMD128 via SIMDe */ +# define inline __inline__ /* circumvent a clang bug */ +# include +# undef inline +# elif defined(__AVX2__) +# include +# elif defined(__SSE2__) +# include +# endif +#endif + +#if defined(_MSC_VER) +# include +#endif + +/* + * One goal of XXH3 is to make it fast on both 32-bit and 64-bit, while + * remaining a true 64-bit/128-bit hash function. + * + * This is done by prioritizing a subset of 64-bit operations that can be + * emulated without too many steps on the average 32-bit machine. + * + * For example, these two lines seem similar, and run equally fast on 64-bit: + * + * xxh_u64 x; + * x ^= (x >> 47); // good + * x ^= (x >> 13); // bad + * + * However, to a 32-bit machine, there is a major difference. + * + * x ^= (x >> 47) looks like this: + * + * x.lo ^= (x.hi >> (47 - 32)); + * + * while x ^= (x >> 13) looks like this: + * + * // note: funnel shifts are not usually cheap. + * x.lo ^= (x.lo >> 13) | (x.hi << (32 - 13)); + * x.hi ^= (x.hi >> 13); + * + * The first one is significantly faster than the second, simply because the + * shift is larger than 32. This means: + * - All the bits we need are in the upper 32 bits, so we can ignore the lower + * 32 bits in the shift. + * - The shift result will always fit in the lower 32 bits, and therefore, + * we can ignore the upper 32 bits in the xor. + * + * Thanks to this optimization, XXH3 only requires these features to be efficient: + * + * - Usable unaligned access + * - A 32-bit or 64-bit ALU + * - If 32-bit, a decent ADC instruction + * - A 32 or 64-bit multiply with a 64-bit result + * - For the 128-bit variant, a decent byteswap helps short inputs. + * + * The first two are already required by XXH32, and almost all 32-bit and 64-bit + * platforms which can run XXH32 can run XXH3 efficiently. + * + * Thumb-1, the classic 16-bit only subset of ARM's instruction set, is one + * notable exception. + * + * First of all, Thumb-1 lacks support for the UMULL instruction which + * performs the important long multiply. This means numerous __aeabi_lmul + * calls. + * + * Second of all, the 8 functional registers are just not enough. + * Setup for __aeabi_lmul, byteshift loads, pointers, and all arithmetic need + * Lo registers, and this shuffling results in thousands more MOVs than A32. + * + * A32 and T32 don't have this limitation. They can access all 14 registers, + * do a 32->64 multiply with UMULL, and the flexible operand allowing free + * shifts is helpful, too. + * + * Therefore, we do a quick sanity check. + * + * If compiling Thumb-1 for a target which supports ARM instructions, we will + * emit a warning, as it is not a "sane" platform to compile for. + * + * Usually, if this happens, it is because of an accident and you probably need + * to specify -march, as you likely meant to compile for a newer architecture. + * + * Credit: large sections of the vectorial and asm source code paths + * have been contributed by @easyaspi314 + */ +#if defined(__thumb__) && !defined(__thumb2__) && defined(__ARM_ARCH_ISA_ARM) +# warning "XXH3 is highly inefficient without ARM or Thumb-2." +#endif + +/* ========================================== + * Vectorization detection + * ========================================== */ + +#ifdef XXH_DOXYGEN +/*! + * @ingroup tuning + * @brief Overrides the vectorization implementation chosen for XXH3. + * + * Can be defined to 0 to disable SIMD or any of the values mentioned in + * @ref XXH_VECTOR_TYPE. + * + * If this is not defined, it uses predefined macros to determine the best + * implementation. + */ +# define XXH_VECTOR XXH_SCALAR +/*! + * @ingroup tuning + * @brief Possible values for @ref XXH_VECTOR. + * + * Note that these are actually implemented as macros. + * + * If this is not defined, it is detected automatically. + * internal macro XXH_X86DISPATCH overrides this. + */ +enum XXH_VECTOR_TYPE /* fake enum */ { + XXH_SCALAR = 0, /*!< Portable scalar version */ + XXH_SSE2 = 1, /*!< + * SSE2 for Pentium 4, Opteron, all x86_64. + * + * @note SSE2 is also guaranteed on Windows 10, macOS, and + * Android x86. + */ + XXH_AVX2 = 2, /*!< AVX2 for Haswell and Bulldozer */ + XXH_AVX512 = 3, /*!< AVX512 for Skylake and Icelake */ + XXH_NEON = 4, /*!< + * NEON for most ARMv7-A, all AArch64, and WASM SIMD128 + * via the SIMDeverywhere polyfill provided with the + * Emscripten SDK. + */ + XXH_VSX = 5, /*!< VSX and ZVector for POWER8/z13 (64-bit) */ + XXH_SVE = 6, /*!< SVE for some ARMv8-A and ARMv9-A */ +}; +/*! + * @ingroup tuning + * @brief Selects the minimum alignment for XXH3's accumulators. + * + * When using SIMD, this should match the alignment required for said vector + * type, so, for example, 32 for AVX2. + * + * Default: Auto detected. + */ +# define XXH_ACC_ALIGN 8 +#endif + +/* Actual definition */ +#ifndef XXH_DOXYGEN +# define XXH_SCALAR 0 +# define XXH_SSE2 1 +# define XXH_AVX2 2 +# define XXH_AVX512 3 +# define XXH_NEON 4 +# define XXH_VSX 5 +# define XXH_SVE 6 +#endif + +#ifndef XXH_VECTOR /* can be defined on command line */ +# if defined(__ARM_FEATURE_SVE) +# define XXH_VECTOR XXH_SVE +# elif ( \ + defined(__ARM_NEON__) || defined(__ARM_NEON) /* gcc */ \ + || defined(_M_ARM) || defined(_M_ARM64) || defined(_M_ARM64EC) /* msvc */ \ + || (defined(__wasm_simd128__) && XXH_HAS_INCLUDE()) /* wasm simd128 via SIMDe */ \ + ) && ( \ + defined(_WIN32) || defined(__LITTLE_ENDIAN__) /* little endian only */ \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) \ + ) +# define XXH_VECTOR XXH_NEON +# elif defined(__AVX512F__) +# define XXH_VECTOR XXH_AVX512 +# elif defined(__AVX2__) +# define XXH_VECTOR XXH_AVX2 +# elif defined(__SSE2__) || defined(_M_AMD64) || defined(_M_X64) || (defined(_M_IX86_FP) && (_M_IX86_FP == 2)) +# define XXH_VECTOR XXH_SSE2 +# elif (defined(__PPC64__) && defined(__POWER8_VECTOR__)) \ + || (defined(__s390x__) && defined(__VEC__)) \ + && defined(__GNUC__) /* TODO: IBM XL */ +# define XXH_VECTOR XXH_VSX +# else +# define XXH_VECTOR XXH_SCALAR +# endif +#endif + +/* __ARM_FEATURE_SVE is only supported by GCC & Clang. */ +#if (XXH_VECTOR == XXH_SVE) && !defined(__ARM_FEATURE_SVE) +# ifdef _MSC_VER +# pragma warning(once : 4606) +# else +# warning "__ARM_FEATURE_SVE isn't supported. Use SCALAR instead." +# endif +# undef XXH_VECTOR +# define XXH_VECTOR XXH_SCALAR +#endif + +/* + * Controls the alignment of the accumulator, + * for compatibility with aligned vector loads, which are usually faster. + */ +#ifndef XXH_ACC_ALIGN +# if defined(XXH_X86DISPATCH) +# define XXH_ACC_ALIGN 64 /* for compatibility with avx512 */ +# elif XXH_VECTOR == XXH_SCALAR /* scalar */ +# define XXH_ACC_ALIGN 8 +# elif XXH_VECTOR == XXH_SSE2 /* sse2 */ +# define XXH_ACC_ALIGN 16 +# elif XXH_VECTOR == XXH_AVX2 /* avx2 */ +# define XXH_ACC_ALIGN 32 +# elif XXH_VECTOR == XXH_NEON /* neon */ +# define XXH_ACC_ALIGN 16 +# elif XXH_VECTOR == XXH_VSX /* vsx */ +# define XXH_ACC_ALIGN 16 +# elif XXH_VECTOR == XXH_AVX512 /* avx512 */ +# define XXH_ACC_ALIGN 64 +# elif XXH_VECTOR == XXH_SVE /* sve */ +# define XXH_ACC_ALIGN 64 +# endif +#endif + +#if defined(XXH_X86DISPATCH) || XXH_VECTOR == XXH_SSE2 \ + || XXH_VECTOR == XXH_AVX2 || XXH_VECTOR == XXH_AVX512 +# define XXH_SEC_ALIGN XXH_ACC_ALIGN +#elif XXH_VECTOR == XXH_SVE +# define XXH_SEC_ALIGN XXH_ACC_ALIGN +#else +# define XXH_SEC_ALIGN 8 +#endif + +#if defined(__GNUC__) || defined(__clang__) +# define XXH_ALIASING __attribute__((__may_alias__)) +#else +# define XXH_ALIASING /* nothing */ +#endif + +/* + * UGLY HACK: + * GCC usually generates the best code with -O3 for xxHash. + * + * However, when targeting AVX2, it is overzealous in its unrolling resulting + * in code roughly 3/4 the speed of Clang. + * + * There are other issues, such as GCC splitting _mm256_loadu_si256 into + * _mm_loadu_si128 + _mm256_inserti128_si256. This is an optimization which + * only applies to Sandy and Ivy Bridge... which don't even support AVX2. + * + * That is why when compiling the AVX2 version, it is recommended to use either + * -O2 -mavx2 -march=haswell + * or + * -O2 -mavx2 -mno-avx256-split-unaligned-load + * for decent performance, or to use Clang instead. + * + * Fortunately, we can control the first one with a pragma that forces GCC into + * -O2, but the other one we can't control without "failed to inline always + * inline function due to target mismatch" warnings. + */ +#if XXH_VECTOR == XXH_AVX2 /* AVX2 */ \ + && defined(__GNUC__) && !defined(__clang__) /* GCC, not Clang */ \ + && defined(__OPTIMIZE__) && XXH_SIZE_OPT <= 0 /* respect -O0 and -Os */ +# pragma GCC push_options +# pragma GCC optimize("-O2") +#endif + +#if XXH_VECTOR == XXH_NEON + +/* + * UGLY HACK: While AArch64 GCC on Linux does not seem to care, on macOS, GCC -O3 + * optimizes out the entire hashLong loop because of the aliasing violation. + * + * However, GCC is also inefficient at load-store optimization with vld1q/vst1q, + * so the only option is to mark it as aliasing. + */ +typedef uint64x2_t xxh_aliasing_uint64x2_t XXH_ALIASING; + +/*! + * @internal + * @brief `vld1q_u64` but faster and alignment-safe. + * + * On AArch64, unaligned access is always safe, but on ARMv7-a, it is only + * *conditionally* safe (`vld1` has an alignment bit like `movdq[ua]` in x86). + * + * GCC for AArch64 sees `vld1q_u8` as an intrinsic instead of a load, so it + * prohibits load-store optimizations. Therefore, a direct dereference is used. + * + * Otherwise, `vld1q_u8` is used with `vreinterpretq_u8_u64` to do a safe + * unaligned load. + */ +#if defined(__aarch64__) && defined(__GNUC__) && !defined(__clang__) +XXH_FORCE_INLINE uint64x2_t XXH_vld1q_u64(void const* ptr) /* silence -Wcast-align */ +{ + return *(xxh_aliasing_uint64x2_t const *)ptr; +} +#else +XXH_FORCE_INLINE uint64x2_t XXH_vld1q_u64(void const* ptr) +{ + return vreinterpretq_u64_u8(vld1q_u8((uint8_t const*)ptr)); +} +#endif + +/*! + * @internal + * @brief `vmlal_u32` on low and high halves of a vector. + * + * This is a workaround for AArch64 GCC < 11 which implemented arm_neon.h with + * inline assembly and were therefore incapable of merging the `vget_{low, high}_u32` + * with `vmlal_u32`. + */ +#if defined(__aarch64__) && defined(__GNUC__) && !defined(__clang__) && __GNUC__ < 11 +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_low_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + /* Inline assembly is the only way */ + __asm__("umlal %0.2d, %1.2s, %2.2s" : "+w" (acc) : "w" (lhs), "w" (rhs)); + return acc; +} +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_high_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + /* This intrinsic works as expected */ + return vmlal_high_u32(acc, lhs, rhs); +} +#else +/* Portable intrinsic versions */ +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_low_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + return vmlal_u32(acc, vget_low_u32(lhs), vget_low_u32(rhs)); +} +/*! @copydoc XXH_vmlal_low_u32 + * Assume the compiler converts this to vmlal_high_u32 on aarch64 */ +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_high_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + return vmlal_u32(acc, vget_high_u32(lhs), vget_high_u32(rhs)); +} +#endif + +/*! + * @ingroup tuning + * @brief Controls the NEON to scalar ratio for XXH3 + * + * This can be set to 2, 4, 6, or 8. + * + * ARM Cortex CPUs are _very_ sensitive to how their pipelines are used. + * + * For example, the Cortex-A73 can dispatch 3 micro-ops per cycle, but only 2 of those + * can be NEON. If you are only using NEON instructions, you are only using 2/3 of the CPU + * bandwidth. + * + * This is even more noticeable on the more advanced cores like the Cortex-A76 which + * can dispatch 8 micro-ops per cycle, but still only 2 NEON micro-ops at once. + * + * Therefore, to make the most out of the pipeline, it is beneficial to run 6 NEON lanes + * and 2 scalar lanes, which is chosen by default. + * + * This does not apply to Apple processors or 32-bit processors, which run better with + * full NEON. These will default to 8. Additionally, size-optimized builds run 8 lanes. + * + * This change benefits CPUs with large micro-op buffers without negatively affecting + * most other CPUs: + * + * | Chipset | Dispatch type | NEON only | 6:2 hybrid | Diff. | + * |:----------------------|:--------------------|----------:|-----------:|------:| + * | Snapdragon 730 (A76) | 2 NEON/8 micro-ops | 8.8 GB/s | 10.1 GB/s | ~16% | + * | Snapdragon 835 (A73) | 2 NEON/3 micro-ops | 5.1 GB/s | 5.3 GB/s | ~5% | + * | Marvell PXA1928 (A53) | In-order dual-issue | 1.9 GB/s | 1.9 GB/s | 0% | + * | Apple M1 | 4 NEON/8 micro-ops | 37.3 GB/s | 36.1 GB/s | ~-3% | + * + * It also seems to fix some bad codegen on GCC, making it almost as fast as clang. + * + * When using WASM SIMD128, if this is 2 or 6, SIMDe will scalarize 2 of the lanes meaning + * it effectively becomes worse 4. + * + * @see XXH3_accumulate_512_neon() + */ +# ifndef XXH3_NEON_LANES +# if (defined(__aarch64__) || defined(__arm64__) || defined(_M_ARM64) || defined(_M_ARM64EC)) \ + && !defined(__APPLE__) && XXH_SIZE_OPT <= 0 +# define XXH3_NEON_LANES 6 +# else +# define XXH3_NEON_LANES XXH_ACC_NB +# endif +# endif +#endif /* XXH_VECTOR == XXH_NEON */ + +/* + * VSX and Z Vector helpers. + * + * This is very messy, and any pull requests to clean this up are welcome. + * + * There are a lot of problems with supporting VSX and s390x, due to + * inconsistent intrinsics, spotty coverage, and multiple endiannesses. + */ +#if XXH_VECTOR == XXH_VSX +/* Annoyingly, these headers _may_ define three macros: `bool`, `vector`, + * and `pixel`. This is a problem for obvious reasons. + * + * These keywords are unnecessary; the spec literally says they are + * equivalent to `__bool`, `__vector`, and `__pixel` and may be undef'd + * after including the header. + * + * We use pragma push_macro/pop_macro to keep the namespace clean. */ +# pragma push_macro("bool") +# pragma push_macro("vector") +# pragma push_macro("pixel") +/* silence potential macro redefined warnings */ +# undef bool +# undef vector +# undef pixel + +# if defined(__s390x__) +# include +# else +# include +# endif + +/* Restore the original macro values, if applicable. */ +# pragma pop_macro("pixel") +# pragma pop_macro("vector") +# pragma pop_macro("bool") + +typedef __vector unsigned long long xxh_u64x2; +typedef __vector unsigned char xxh_u8x16; +typedef __vector unsigned xxh_u32x4; + +/* + * UGLY HACK: Similar to aarch64 macOS GCC, s390x GCC has the same aliasing issue. + */ +typedef xxh_u64x2 xxh_aliasing_u64x2 XXH_ALIASING; + +# ifndef XXH_VSX_BE +# if defined(__BIG_ENDIAN__) \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) +# define XXH_VSX_BE 1 +# elif defined(__VEC_ELEMENT_REG_ORDER__) && __VEC_ELEMENT_REG_ORDER__ == __ORDER_BIG_ENDIAN__ +# warning "-maltivec=be is not recommended. Please use native endianness." +# define XXH_VSX_BE 1 +# else +# define XXH_VSX_BE 0 +# endif +# endif /* !defined(XXH_VSX_BE) */ + +# if XXH_VSX_BE +# if defined(__POWER9_VECTOR__) || (defined(__clang__) && defined(__s390x__)) +# define XXH_vec_revb vec_revb +# else +/*! + * A polyfill for POWER9's vec_revb(). + */ +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_revb(xxh_u64x2 val) +{ + xxh_u8x16 const vByteSwap = { 0x07, 0x06, 0x05, 0x04, 0x03, 0x02, 0x01, 0x00, + 0x0F, 0x0E, 0x0D, 0x0C, 0x0B, 0x0A, 0x09, 0x08 }; + return vec_perm(val, val, vByteSwap); +} +# endif +# endif /* XXH_VSX_BE */ + +/*! + * Performs an unaligned vector load and byte swaps it on big endian. + */ +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_loadu(const void *ptr) +{ + xxh_u64x2 ret; + XXH_memcpy(&ret, ptr, sizeof(xxh_u64x2)); +# if XXH_VSX_BE + ret = XXH_vec_revb(ret); +# endif + return ret; +} + +/* + * vec_mulo and vec_mule are very problematic intrinsics on PowerPC + * + * These intrinsics weren't added until GCC 8, despite existing for a while, + * and they are endian dependent. Also, their meaning swap depending on version. + * */ +# if defined(__s390x__) + /* s390x is always big endian, no issue on this platform */ +# define XXH_vec_mulo vec_mulo +# define XXH_vec_mule vec_mule +# elif defined(__clang__) && XXH_HAS_BUILTIN(__builtin_altivec_vmuleuw) && !defined(__ibmxl__) +/* Clang has a better way to control this, we can just use the builtin which doesn't swap. */ + /* The IBM XL Compiler (which defined __clang__) only implements the vec_* operations */ +# define XXH_vec_mulo __builtin_altivec_vmulouw +# define XXH_vec_mule __builtin_altivec_vmuleuw +# else +/* gcc needs inline assembly */ +/* Adapted from https://github.com/google/highwayhash/blob/master/highwayhash/hh_vsx.h. */ +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_mulo(xxh_u32x4 a, xxh_u32x4 b) +{ + xxh_u64x2 result; + __asm__("vmulouw %0, %1, %2" : "=v" (result) : "v" (a), "v" (b)); + return result; +} +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_mule(xxh_u32x4 a, xxh_u32x4 b) +{ + xxh_u64x2 result; + __asm__("vmuleuw %0, %1, %2" : "=v" (result) : "v" (a), "v" (b)); + return result; +} +# endif /* XXH_vec_mulo, XXH_vec_mule */ +#endif /* XXH_VECTOR == XXH_VSX */ + +#if XXH_VECTOR == XXH_SVE +#define ACCRND(acc, offset) \ +do { \ + svuint64_t input_vec = svld1_u64(mask, xinput + offset); \ + svuint64_t secret_vec = svld1_u64(mask, xsecret + offset); \ + svuint64_t mixed = sveor_u64_x(mask, secret_vec, input_vec); \ + svuint64_t swapped = svtbl_u64(input_vec, kSwap); \ + svuint64_t mixed_lo = svextw_u64_x(mask, mixed); \ + svuint64_t mixed_hi = svlsr_n_u64_x(mask, mixed, 32); \ + svuint64_t mul = svmad_u64_x(mask, mixed_lo, mixed_hi, swapped); \ + acc = svadd_u64_x(mask, acc, mul); \ +} while (0) +#endif /* XXH_VECTOR == XXH_SVE */ + +/* prefetch + * can be disabled, by declaring XXH_NO_PREFETCH build macro */ +#if defined(XXH_NO_PREFETCH) +# define XXH_PREFETCH(ptr) (void)(ptr) /* disabled */ +#else +# if XXH_SIZE_OPT >= 1 +# define XXH_PREFETCH(ptr) (void)(ptr) +# elif defined(_MSC_VER) && (defined(_M_X64) || defined(_M_IX86)) /* _mm_prefetch() not defined outside of x86/x64 */ +# include /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */ +# define XXH_PREFETCH(ptr) _mm_prefetch((const char*)(ptr), _MM_HINT_T0) +# elif defined(__GNUC__) && ( (__GNUC__ >= 4) || ( (__GNUC__ == 3) && (__GNUC_MINOR__ >= 1) ) ) +# define XXH_PREFETCH(ptr) __builtin_prefetch((ptr), 0 /* rw==read */, 3 /* locality */) +# else +# define XXH_PREFETCH(ptr) (void)(ptr) /* disabled */ +# endif +#endif /* XXH_NO_PREFETCH */ + + +/* ========================================== + * XXH3 default settings + * ========================================== */ + +#define XXH_SECRET_DEFAULT_SIZE 192 /* minimum XXH3_SECRET_SIZE_MIN */ + +#if (XXH_SECRET_DEFAULT_SIZE < XXH3_SECRET_SIZE_MIN) +# error "default keyset is not large enough" +#endif + +/*! Pseudorandom secret taken directly from FARSH. */ +XXH_ALIGN(64) static const xxh_u8 XXH3_kSecret[XXH_SECRET_DEFAULT_SIZE] = { + 0xb8, 0xfe, 0x6c, 0x39, 0x23, 0xa4, 0x4b, 0xbe, 0x7c, 0x01, 0x81, 0x2c, 0xf7, 0x21, 0xad, 0x1c, + 0xde, 0xd4, 0x6d, 0xe9, 0x83, 0x90, 0x97, 0xdb, 0x72, 0x40, 0xa4, 0xa4, 0xb7, 0xb3, 0x67, 0x1f, + 0xcb, 0x79, 0xe6, 0x4e, 0xcc, 0xc0, 0xe5, 0x78, 0x82, 0x5a, 0xd0, 0x7d, 0xcc, 0xff, 0x72, 0x21, + 0xb8, 0x08, 0x46, 0x74, 0xf7, 0x43, 0x24, 0x8e, 0xe0, 0x35, 0x90, 0xe6, 0x81, 0x3a, 0x26, 0x4c, + 0x3c, 0x28, 0x52, 0xbb, 0x91, 0xc3, 0x00, 0xcb, 0x88, 0xd0, 0x65, 0x8b, 0x1b, 0x53, 0x2e, 0xa3, + 0x71, 0x64, 0x48, 0x97, 0xa2, 0x0d, 0xf9, 0x4e, 0x38, 0x19, 0xef, 0x46, 0xa9, 0xde, 0xac, 0xd8, + 0xa8, 0xfa, 0x76, 0x3f, 0xe3, 0x9c, 0x34, 0x3f, 0xf9, 0xdc, 0xbb, 0xc7, 0xc7, 0x0b, 0x4f, 0x1d, + 0x8a, 0x51, 0xe0, 0x4b, 0xcd, 0xb4, 0x59, 0x31, 0xc8, 0x9f, 0x7e, 0xc9, 0xd9, 0x78, 0x73, 0x64, + 0xea, 0xc5, 0xac, 0x83, 0x34, 0xd3, 0xeb, 0xc3, 0xc5, 0x81, 0xa0, 0xff, 0xfa, 0x13, 0x63, 0xeb, + 0x17, 0x0d, 0xdd, 0x51, 0xb7, 0xf0, 0xda, 0x49, 0xd3, 0x16, 0x55, 0x26, 0x29, 0xd4, 0x68, 0x9e, + 0x2b, 0x16, 0xbe, 0x58, 0x7d, 0x47, 0xa1, 0xfc, 0x8f, 0xf8, 0xb8, 0xd1, 0x7a, 0xd0, 0x31, 0xce, + 0x45, 0xcb, 0x3a, 0x8f, 0x95, 0x16, 0x04, 0x28, 0xaf, 0xd7, 0xfb, 0xca, 0xbb, 0x4b, 0x40, 0x7e, +}; + +static const xxh_u64 PRIME_MX1 = 0x165667919E3779F9ULL; /*!< 0b0001011001010110011001111001000110011110001101110111100111111001 */ +static const xxh_u64 PRIME_MX2 = 0x9FB21C651E98DF25ULL; /*!< 0b1001111110110010000111000110010100011110100110001101111100100101 */ + +#ifdef XXH_OLD_NAMES +# define kSecret XXH3_kSecret +#endif + +#ifdef XXH_DOXYGEN +/*! + * @brief Calculates a 32-bit to 64-bit long multiply. + * + * Implemented as a macro. + * + * Wraps `__emulu` on MSVC x86 because it tends to call `__allmul` when it doesn't + * need to (but it shouldn't need to anyways, it is about 7 instructions to do + * a 64x64 multiply...). Since we know that this will _always_ emit `MULL`, we + * use that instead of the normal method. + * + * If you are compiling for platforms like Thumb-1 and don't have a better option, + * you may also want to write your own long multiply routine here. + * + * @param x, y Numbers to be multiplied + * @return 64-bit product of the low 32 bits of @p x and @p y. + */ +XXH_FORCE_INLINE xxh_u64 +XXH_mult32to64(xxh_u64 x, xxh_u64 y) +{ + return (x & 0xFFFFFFFF) * (y & 0xFFFFFFFF); +} +#elif defined(_MSC_VER) && defined(_M_IX86) +# define XXH_mult32to64(x, y) __emulu((unsigned)(x), (unsigned)(y)) +#else +/* + * Downcast + upcast is usually better than masking on older compilers like + * GCC 4.2 (especially 32-bit ones), all without affecting newer compilers. + * + * The other method, (x & 0xFFFFFFFF) * (y & 0xFFFFFFFF), will AND both operands + * and perform a full 64x64 multiply -- entirely redundant on 32-bit. + */ +# define XXH_mult32to64(x, y) ((xxh_u64)(xxh_u32)(x) * (xxh_u64)(xxh_u32)(y)) +#endif + +/*! + * @brief Calculates a 64->128-bit long multiply. + * + * Uses `__uint128_t` and `_umul128` if available, otherwise uses a scalar + * version. + * + * @param lhs , rhs The 64-bit integers to be multiplied + * @return The 128-bit result represented in an @ref XXH128_hash_t. + */ +static XXH128_hash_t +XXH_mult64to128(xxh_u64 lhs, xxh_u64 rhs) +{ + /* + * GCC/Clang __uint128_t method. + * + * On most 64-bit targets, GCC and Clang define a __uint128_t type. + * This is usually the best way as it usually uses a native long 64-bit + * multiply, such as MULQ on x86_64 or MUL + UMULH on aarch64. + * + * Usually. + * + * Despite being a 32-bit platform, Clang (and emscripten) define this type + * despite not having the arithmetic for it. This results in a laggy + * compiler builtin call which calculates a full 128-bit multiply. + * In that case it is best to use the portable one. + * https://github.com/Cyan4973/xxHash/issues/211#issuecomment-515575677 + */ +#if (defined(__GNUC__) || defined(__clang__)) && !defined(__wasm__) \ + && defined(__SIZEOF_INT128__) \ + || (defined(_INTEGRAL_MAX_BITS) && _INTEGRAL_MAX_BITS >= 128) + + __uint128_t const product = (__uint128_t)lhs * (__uint128_t)rhs; + XXH128_hash_t r128; + r128.low64 = (xxh_u64)(product); + r128.high64 = (xxh_u64)(product >> 64); + return r128; + + /* + * MSVC for x64's _umul128 method. + * + * xxh_u64 _umul128(xxh_u64 Multiplier, xxh_u64 Multiplicand, xxh_u64 *HighProduct); + * + * This compiles to single operand MUL on x64. + */ +#elif (defined(_M_X64) || defined(_M_IA64)) && !defined(_M_ARM64EC) + +#ifndef _MSC_VER +# pragma intrinsic(_umul128) +#endif + xxh_u64 product_high; + xxh_u64 const product_low = _umul128(lhs, rhs, &product_high); + XXH128_hash_t r128; + r128.low64 = product_low; + r128.high64 = product_high; + return r128; + + /* + * MSVC for ARM64's __umulh method. + * + * This compiles to the same MUL + UMULH as GCC/Clang's __uint128_t method. + */ +#elif defined(_M_ARM64) || defined(_M_ARM64EC) + +#ifndef _MSC_VER +# pragma intrinsic(__umulh) +#endif + XXH128_hash_t r128; + r128.low64 = lhs * rhs; + r128.high64 = __umulh(lhs, rhs); + return r128; + +#else + /* + * Portable scalar method. Optimized for 32-bit and 64-bit ALUs. + * + * This is a fast and simple grade school multiply, which is shown below + * with base 10 arithmetic instead of base 0x100000000. + * + * 9 3 // D2 lhs = 93 + * x 7 5 // D2 rhs = 75 + * ---------- + * 1 5 // D2 lo_lo = (93 % 10) * (75 % 10) = 15 + * 4 5 | // D2 hi_lo = (93 / 10) * (75 % 10) = 45 + * 2 1 | // D2 lo_hi = (93 % 10) * (75 / 10) = 21 + * + 6 3 | | // D2 hi_hi = (93 / 10) * (75 / 10) = 63 + * --------- + * 2 7 | // D2 cross = (15 / 10) + (45 % 10) + 21 = 27 + * + 6 7 | | // D2 upper = (27 / 10) + (45 / 10) + 63 = 67 + * --------- + * 6 9 7 5 // D4 res = (27 * 10) + (15 % 10) + (67 * 100) = 6975 + * + * The reasons for adding the products like this are: + * 1. It avoids manual carry tracking. Just like how + * (9 * 9) + 9 + 9 = 99, the same applies with this for UINT64_MAX. + * This avoids a lot of complexity. + * + * 2. It hints for, and on Clang, compiles to, the powerful UMAAL + * instruction available in ARM's Digital Signal Processing extension + * in 32-bit ARMv6 and later, which is shown below: + * + * void UMAAL(xxh_u32 *RdLo, xxh_u32 *RdHi, xxh_u32 Rn, xxh_u32 Rm) + * { + * xxh_u64 product = (xxh_u64)*RdLo * (xxh_u64)*RdHi + Rn + Rm; + * *RdLo = (xxh_u32)(product & 0xFFFFFFFF); + * *RdHi = (xxh_u32)(product >> 32); + * } + * + * This instruction was designed for efficient long multiplication, and + * allows this to be calculated in only 4 instructions at speeds + * comparable to some 64-bit ALUs. + * + * 3. It isn't terrible on other platforms. Usually this will be a couple + * of 32-bit ADD/ADCs. + */ + + /* First calculate all of the cross products. */ + xxh_u64 const lo_lo = XXH_mult32to64(lhs & 0xFFFFFFFF, rhs & 0xFFFFFFFF); + xxh_u64 const hi_lo = XXH_mult32to64(lhs >> 32, rhs & 0xFFFFFFFF); + xxh_u64 const lo_hi = XXH_mult32to64(lhs & 0xFFFFFFFF, rhs >> 32); + xxh_u64 const hi_hi = XXH_mult32to64(lhs >> 32, rhs >> 32); + + /* Now add the products together. These will never overflow. */ + xxh_u64 const cross = (lo_lo >> 32) + (hi_lo & 0xFFFFFFFF) + lo_hi; + xxh_u64 const upper = (hi_lo >> 32) + (cross >> 32) + hi_hi; + xxh_u64 const lower = (cross << 32) | (lo_lo & 0xFFFFFFFF); + + XXH128_hash_t r128; + r128.low64 = lower; + r128.high64 = upper; + return r128; +#endif +} + +/*! + * @brief Calculates a 64-bit to 128-bit multiply, then XOR folds it. + * + * The reason for the separate function is to prevent passing too many structs + * around by value. This will hopefully inline the multiply, but we don't force it. + * + * @param lhs , rhs The 64-bit integers to multiply + * @return The low 64 bits of the product XOR'd by the high 64 bits. + * @see XXH_mult64to128() + */ +static xxh_u64 +XXH3_mul128_fold64(xxh_u64 lhs, xxh_u64 rhs) +{ + XXH128_hash_t product = XXH_mult64to128(lhs, rhs); + return product.low64 ^ product.high64; +} + +/*! Seems to produce slightly better code on GCC for some reason. */ +XXH_FORCE_INLINE XXH_CONSTF xxh_u64 XXH_xorshift64(xxh_u64 v64, int shift) +{ + XXH_ASSERT(0 <= shift && shift < 64); + return v64 ^ (v64 >> shift); +} + +/* + * This is a fast avalanche stage, + * suitable when input bits are already partially mixed + */ +static XXH64_hash_t XXH3_avalanche(xxh_u64 h64) +{ + h64 = XXH_xorshift64(h64, 37); + h64 *= PRIME_MX1; + h64 = XXH_xorshift64(h64, 32); + return h64; +} + +/* + * This is a stronger avalanche, + * inspired by Pelle Evensen's rrmxmx + * preferable when input has not been previously mixed + */ +static XXH64_hash_t XXH3_rrmxmx(xxh_u64 h64, xxh_u64 len) +{ + /* this mix is inspired by Pelle Evensen's rrmxmx */ + h64 ^= XXH_rotl64(h64, 49) ^ XXH_rotl64(h64, 24); + h64 *= PRIME_MX2; + h64 ^= (h64 >> 35) + len ; + h64 *= PRIME_MX2; + return XXH_xorshift64(h64, 28); +} + + +/* ========================================== + * Short keys + * ========================================== + * One of the shortcomings of XXH32 and XXH64 was that their performance was + * sub-optimal on short lengths. It used an iterative algorithm which strongly + * favored lengths that were a multiple of 4 or 8. + * + * Instead of iterating over individual inputs, we use a set of single shot + * functions which piece together a range of lengths and operate in constant time. + * + * Additionally, the number of multiplies has been significantly reduced. This + * reduces latency, especially when emulating 64-bit multiplies on 32-bit. + * + * Depending on the platform, this may or may not be faster than XXH32, but it + * is almost guaranteed to be faster than XXH64. + */ + +/* + * At very short lengths, there isn't enough input to fully hide secrets, or use + * the entire secret. + * + * There is also only a limited amount of mixing we can do before significantly + * impacting performance. + * + * Therefore, we use different sections of the secret and always mix two secret + * samples with an XOR. This should have no effect on performance on the + * seedless or withSeed variants because everything _should_ be constant folded + * by modern compilers. + * + * The XOR mixing hides individual parts of the secret and increases entropy. + * + * This adds an extra layer of strength for custom secrets. + */ +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_1to3_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(1 <= len && len <= 3); + XXH_ASSERT(secret != NULL); + /* + * len = 1: combined = { input[0], 0x01, input[0], input[0] } + * len = 2: combined = { input[1], 0x02, input[0], input[1] } + * len = 3: combined = { input[2], 0x03, input[0], input[1] } + */ + { xxh_u8 const c1 = input[0]; + xxh_u8 const c2 = input[len >> 1]; + xxh_u8 const c3 = input[len - 1]; + xxh_u32 const combined = ((xxh_u32)c1 << 16) | ((xxh_u32)c2 << 24) + | ((xxh_u32)c3 << 0) | ((xxh_u32)len << 8); + xxh_u64 const bitflip = (XXH_readLE32(secret) ^ XXH_readLE32(secret+4)) + seed; + xxh_u64 const keyed = (xxh_u64)combined ^ bitflip; + return XXH64_avalanche(keyed); + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_4to8_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(4 <= len && len <= 8); + seed ^= (xxh_u64)XXH_swap32((xxh_u32)seed) << 32; + { xxh_u32 const input1 = XXH_readLE32(input); + xxh_u32 const input2 = XXH_readLE32(input + len - 4); + xxh_u64 const bitflip = (XXH_readLE64(secret+8) ^ XXH_readLE64(secret+16)) - seed; + xxh_u64 const input64 = input2 + (((xxh_u64)input1) << 32); + xxh_u64 const keyed = input64 ^ bitflip; + return XXH3_rrmxmx(keyed, len); + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_9to16_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(9 <= len && len <= 16); + { xxh_u64 const bitflip1 = (XXH_readLE64(secret+24) ^ XXH_readLE64(secret+32)) + seed; + xxh_u64 const bitflip2 = (XXH_readLE64(secret+40) ^ XXH_readLE64(secret+48)) - seed; + xxh_u64 const input_lo = XXH_readLE64(input) ^ bitflip1; + xxh_u64 const input_hi = XXH_readLE64(input + len - 8) ^ bitflip2; + xxh_u64 const acc = len + + XXH_swap64(input_lo) + input_hi + + XXH3_mul128_fold64(input_lo, input_hi); + return XXH3_avalanche(acc); + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_0to16_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(len <= 16); + { if (XXH_likely(len > 8)) return XXH3_len_9to16_64b(input, len, secret, seed); + if (XXH_likely(len >= 4)) return XXH3_len_4to8_64b(input, len, secret, seed); + if (len) return XXH3_len_1to3_64b(input, len, secret, seed); + return XXH64_avalanche(seed ^ (XXH_readLE64(secret+56) ^ XXH_readLE64(secret+64))); + } +} + +/* + * DISCLAIMER: There are known *seed-dependent* multicollisions here due to + * multiplication by zero, affecting hashes of lengths 17 to 240. + * + * However, they are very unlikely. + * + * Keep this in mind when using the unseeded XXH3_64bits() variant: As with all + * unseeded non-cryptographic hashes, it does not attempt to defend itself + * against specially crafted inputs, only random inputs. + * + * Compared to classic UMAC where a 1 in 2^31 chance of 4 consecutive bytes + * cancelling out the secret is taken an arbitrary number of times (addressed + * in XXH3_accumulate_512), this collision is very unlikely with random inputs + * and/or proper seeding: + * + * This only has a 1 in 2^63 chance of 8 consecutive bytes cancelling out, in a + * function that is only called up to 16 times per hash with up to 240 bytes of + * input. + * + * This is not too bad for a non-cryptographic hash function, especially with + * only 64 bit outputs. + * + * The 128-bit variant (which trades some speed for strength) is NOT affected + * by this, although it is always a good idea to use a proper seed if you care + * about strength. + */ +XXH_FORCE_INLINE xxh_u64 XXH3_mix16B(const xxh_u8* XXH_RESTRICT input, + const xxh_u8* XXH_RESTRICT secret, xxh_u64 seed64) +{ +#if defined(__GNUC__) && !defined(__clang__) /* GCC, not Clang */ \ + && defined(__i386__) && defined(__SSE2__) /* x86 + SSE2 */ \ + && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable like XXH32 hack */ + /* + * UGLY HACK: + * GCC for x86 tends to autovectorize the 128-bit multiply, resulting in + * slower code. + * + * By forcing seed64 into a register, we disrupt the cost model and + * cause it to scalarize. See `XXH32_round()` + * + * FIXME: Clang's output is still _much_ faster -- On an AMD Ryzen 3600, + * XXH3_64bits @ len=240 runs at 4.6 GB/s with Clang 9, but 3.3 GB/s on + * GCC 9.2, despite both emitting scalar code. + * + * GCC generates much better scalar code than Clang for the rest of XXH3, + * which is why finding a more optimal codepath is an interest. + */ + XXH_COMPILER_GUARD(seed64); +#endif + { xxh_u64 const input_lo = XXH_readLE64(input); + xxh_u64 const input_hi = XXH_readLE64(input+8); + return XXH3_mul128_fold64( + input_lo ^ (XXH_readLE64(secret) + seed64), + input_hi ^ (XXH_readLE64(secret+8) - seed64) + ); + } +} + +/* For mid range keys, XXH3 uses a Mum-hash variant. */ +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_17to128_64b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(16 < len && len <= 128); + + { xxh_u64 acc = len * XXH_PRIME64_1; +#if XXH_SIZE_OPT >= 1 + /* Smaller and cleaner, but slightly slower. */ + unsigned int i = (unsigned int)(len - 1) / 32; + do { + acc += XXH3_mix16B(input+16 * i, secret+32*i, seed); + acc += XXH3_mix16B(input+len-16*(i+1), secret+32*i+16, seed); + } while (i-- != 0); +#else + if (len > 32) { + if (len > 64) { + if (len > 96) { + acc += XXH3_mix16B(input+48, secret+96, seed); + acc += XXH3_mix16B(input+len-64, secret+112, seed); + } + acc += XXH3_mix16B(input+32, secret+64, seed); + acc += XXH3_mix16B(input+len-48, secret+80, seed); + } + acc += XXH3_mix16B(input+16, secret+32, seed); + acc += XXH3_mix16B(input+len-32, secret+48, seed); + } + acc += XXH3_mix16B(input+0, secret+0, seed); + acc += XXH3_mix16B(input+len-16, secret+16, seed); +#endif + return XXH3_avalanche(acc); + } +} + +XXH_NO_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_129to240_64b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); + + #define XXH3_MIDSIZE_STARTOFFSET 3 + #define XXH3_MIDSIZE_LASTOFFSET 17 + + { xxh_u64 acc = len * XXH_PRIME64_1; + xxh_u64 acc_end; + unsigned int const nbRounds = (unsigned int)len / 16; + unsigned int i; + XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); + for (i=0; i<8; i++) { + acc += XXH3_mix16B(input+(16*i), secret+(16*i), seed); + } + /* last bytes */ + acc_end = XXH3_mix16B(input + len - 16, secret + XXH3_SECRET_SIZE_MIN - XXH3_MIDSIZE_LASTOFFSET, seed); + XXH_ASSERT(nbRounds >= 8); + acc = XXH3_avalanche(acc); +#if defined(__clang__) /* Clang */ \ + && (defined(__ARM_NEON) || defined(__ARM_NEON__)) /* NEON */ \ + && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable */ + /* + * UGLY HACK: + * Clang for ARMv7-A tries to vectorize this loop, similar to GCC x86. + * In everywhere else, it uses scalar code. + * + * For 64->128-bit multiplies, even if the NEON was 100% optimal, it + * would still be slower than UMAAL (see XXH_mult64to128). + * + * Unfortunately, Clang doesn't handle the long multiplies properly and + * converts them to the nonexistent "vmulq_u64" intrinsic, which is then + * scalarized into an ugly mess of VMOV.32 instructions. + * + * This mess is difficult to avoid without turning autovectorization + * off completely, but they are usually relatively minor and/or not + * worth it to fix. + * + * This loop is the easiest to fix, as unlike XXH32, this pragma + * _actually works_ because it is a loop vectorization instead of an + * SLP vectorization. + */ + #pragma clang loop vectorize(disable) +#endif + for (i=8 ; i < nbRounds; i++) { + /* + * Prevents clang for unrolling the acc loop and interleaving with this one. + */ + XXH_COMPILER_GUARD(acc); + acc_end += XXH3_mix16B(input+(16*i), secret+(16*(i-8)) + XXH3_MIDSIZE_STARTOFFSET, seed); + } + return XXH3_avalanche(acc + acc_end); + } +} + + +/* ======= Long Keys ======= */ + +#define XXH_STRIPE_LEN 64 +#define XXH_SECRET_CONSUME_RATE 8 /* nb of secret bytes consumed at each accumulation */ +#define XXH_ACC_NB (XXH_STRIPE_LEN / sizeof(xxh_u64)) + +#ifdef XXH_OLD_NAMES +# define STRIPE_LEN XXH_STRIPE_LEN +# define ACC_NB XXH_ACC_NB +#endif + +#ifndef XXH_PREFETCH_DIST +# ifdef __clang__ +# define XXH_PREFETCH_DIST 320 +# else +# if (XXH_VECTOR == XXH_AVX512) +# define XXH_PREFETCH_DIST 512 +# else +# define XXH_PREFETCH_DIST 384 +# endif +# endif /* __clang__ */ +#endif /* XXH_PREFETCH_DIST */ + +/* + * These macros are to generate an XXH3_accumulate() function. + * The two arguments select the name suffix and target attribute. + * + * The name of this symbol is XXH3_accumulate_() and it calls + * XXH3_accumulate_512_(). + * + * It may be useful to hand implement this function if the compiler fails to + * optimize the inline function. + */ +#define XXH3_ACCUMULATE_TEMPLATE(name) \ +void \ +XXH3_accumulate_##name(xxh_u64* XXH_RESTRICT acc, \ + const xxh_u8* XXH_RESTRICT input, \ + const xxh_u8* XXH_RESTRICT secret, \ + size_t nbStripes) \ +{ \ + size_t n; \ + for (n = 0; n < nbStripes; n++ ) { \ + const xxh_u8* const in = input + n*XXH_STRIPE_LEN; \ + XXH_PREFETCH(in + XXH_PREFETCH_DIST); \ + XXH3_accumulate_512_##name( \ + acc, \ + in, \ + secret + n*XXH_SECRET_CONSUME_RATE); \ + } \ +} + + +XXH_FORCE_INLINE void XXH_writeLE64(void* dst, xxh_u64 v64) +{ + if (!XXH_CPU_LITTLE_ENDIAN) v64 = XXH_swap64(v64); + XXH_memcpy(dst, &v64, sizeof(v64)); +} + +/* Several intrinsic functions below are supposed to accept __int64 as argument, + * as documented in https://software.intel.com/sites/landingpage/IntrinsicsGuide/ . + * However, several environments do not define __int64 type, + * requiring a workaround. + */ +#if !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) + typedef int64_t xxh_i64; +#else + /* the following type must have a width of 64-bit */ + typedef long long xxh_i64; +#endif + + +/* + * XXH3_accumulate_512 is the tightest loop for long inputs, and it is the most optimized. + * + * It is a hardened version of UMAC, based off of FARSH's implementation. + * + * This was chosen because it adapts quite well to 32-bit, 64-bit, and SIMD + * implementations, and it is ridiculously fast. + * + * We harden it by mixing the original input to the accumulators as well as the product. + * + * This means that in the (relatively likely) case of a multiply by zero, the + * original input is preserved. + * + * On 128-bit inputs, we swap 64-bit pairs when we add the input to improve + * cross-pollination, as otherwise the upper and lower halves would be + * essentially independent. + * + * This doesn't matter on 64-bit hashes since they all get merged together in + * the end, so we skip the extra step. + * + * Both XXH3_64bits and XXH3_128bits use this subroutine. + */ + +#if (XXH_VECTOR == XXH_AVX512) \ + || (defined(XXH_DISPATCH_AVX512) && XXH_DISPATCH_AVX512 != 0) + +#ifndef XXH_TARGET_AVX512 +# define XXH_TARGET_AVX512 /* disable attribute target */ +#endif + +XXH_FORCE_INLINE XXH_TARGET_AVX512 void +XXH3_accumulate_512_avx512(void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + __m512i* const xacc = (__m512i *) acc; + XXH_ASSERT((((size_t)acc) & 63) == 0); + XXH_STATIC_ASSERT(XXH_STRIPE_LEN == sizeof(__m512i)); + + { + /* data_vec = input[0]; */ + __m512i const data_vec = _mm512_loadu_si512 (input); + /* key_vec = secret[0]; */ + __m512i const key_vec = _mm512_loadu_si512 (secret); + /* data_key = data_vec ^ key_vec; */ + __m512i const data_key = _mm512_xor_si512 (data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m512i const data_key_lo = _mm512_srli_epi64 (data_key, 32); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m512i const product = _mm512_mul_epu32 (data_key, data_key_lo); + /* xacc[0] += swap(data_vec); */ + __m512i const data_swap = _mm512_shuffle_epi32(data_vec, (_MM_PERM_ENUM)_MM_SHUFFLE(1, 0, 3, 2)); + __m512i const sum = _mm512_add_epi64(*xacc, data_swap); + /* xacc[0] += product; */ + *xacc = _mm512_add_epi64(product, sum); + } +} +XXH_FORCE_INLINE XXH_TARGET_AVX512 XXH3_ACCUMULATE_TEMPLATE(avx512) + +/* + * XXH3_scrambleAcc: Scrambles the accumulators to improve mixing. + * + * Multiplication isn't perfect, as explained by Google in HighwayHash: + * + * // Multiplication mixes/scrambles bytes 0-7 of the 64-bit result to + * // varying degrees. In descending order of goodness, bytes + * // 3 4 2 5 1 6 0 7 have quality 228 224 164 160 100 96 36 32. + * // As expected, the upper and lower bytes are much worse. + * + * Source: https://github.com/google/highwayhash/blob/0aaf66b/highwayhash/hh_avx2.h#L291 + * + * Since our algorithm uses a pseudorandom secret to add some variance into the + * mix, we don't need to (or want to) mix as often or as much as HighwayHash does. + * + * This isn't as tight as XXH3_accumulate, but still written in SIMD to avoid + * extraction. + * + * Both XXH3_64bits and XXH3_128bits use this subroutine. + */ + +XXH_FORCE_INLINE XXH_TARGET_AVX512 void +XXH3_scrambleAcc_avx512(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 63) == 0); + XXH_STATIC_ASSERT(XXH_STRIPE_LEN == sizeof(__m512i)); + { __m512i* const xacc = (__m512i*) acc; + const __m512i prime32 = _mm512_set1_epi32((int)XXH_PRIME32_1); + + /* xacc[0] ^= (xacc[0] >> 47) */ + __m512i const acc_vec = *xacc; + __m512i const shifted = _mm512_srli_epi64 (acc_vec, 47); + /* xacc[0] ^= secret; */ + __m512i const key_vec = _mm512_loadu_si512 (secret); + __m512i const data_key = _mm512_ternarylogic_epi32(key_vec, acc_vec, shifted, 0x96 /* key_vec ^ acc_vec ^ shifted */); + + /* xacc[0] *= XXH_PRIME32_1; */ + __m512i const data_key_hi = _mm512_srli_epi64 (data_key, 32); + __m512i const prod_lo = _mm512_mul_epu32 (data_key, prime32); + __m512i const prod_hi = _mm512_mul_epu32 (data_key_hi, prime32); + *xacc = _mm512_add_epi64(prod_lo, _mm512_slli_epi64(prod_hi, 32)); + } +} + +XXH_FORCE_INLINE XXH_TARGET_AVX512 void +XXH3_initCustomSecret_avx512(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 63) == 0); + XXH_STATIC_ASSERT(XXH_SEC_ALIGN == 64); + XXH_ASSERT(((size_t)customSecret & 63) == 0); + (void)(&XXH_writeLE64); + { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / sizeof(__m512i); + __m512i const seed_pos = _mm512_set1_epi64((xxh_i64)seed64); + __m512i const seed = _mm512_mask_sub_epi64(seed_pos, 0xAA, _mm512_set1_epi8(0), seed_pos); + + const __m512i* const src = (const __m512i*) ((const void*) XXH3_kSecret); + __m512i* const dest = ( __m512i*) customSecret; + int i; + XXH_ASSERT(((size_t)src & 63) == 0); /* control alignment */ + XXH_ASSERT(((size_t)dest & 63) == 0); + for (i=0; i < nbRounds; ++i) { + dest[i] = _mm512_add_epi64(_mm512_load_si512(src + i), seed); + } } +} + +#endif + +#if (XXH_VECTOR == XXH_AVX2) \ + || (defined(XXH_DISPATCH_AVX2) && XXH_DISPATCH_AVX2 != 0) + +#ifndef XXH_TARGET_AVX2 +# define XXH_TARGET_AVX2 /* disable attribute target */ +#endif + +XXH_FORCE_INLINE XXH_TARGET_AVX2 void +XXH3_accumulate_512_avx2( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 31) == 0); + { __m256i* const xacc = (__m256i *) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ + const __m256i* const xinput = (const __m256i *) input; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ + const __m256i* const xsecret = (const __m256i *) secret; + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m256i); i++) { + /* data_vec = xinput[i]; */ + __m256i const data_vec = _mm256_loadu_si256 (xinput+i); + /* key_vec = xsecret[i]; */ + __m256i const key_vec = _mm256_loadu_si256 (xsecret+i); + /* data_key = data_vec ^ key_vec; */ + __m256i const data_key = _mm256_xor_si256 (data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m256i const data_key_lo = _mm256_srli_epi64 (data_key, 32); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m256i const product = _mm256_mul_epu32 (data_key, data_key_lo); + /* xacc[i] += swap(data_vec); */ + __m256i const data_swap = _mm256_shuffle_epi32(data_vec, _MM_SHUFFLE(1, 0, 3, 2)); + __m256i const sum = _mm256_add_epi64(xacc[i], data_swap); + /* xacc[i] += product; */ + xacc[i] = _mm256_add_epi64(product, sum); + } } +} +XXH_FORCE_INLINE XXH_TARGET_AVX2 XXH3_ACCUMULATE_TEMPLATE(avx2) + +XXH_FORCE_INLINE XXH_TARGET_AVX2 void +XXH3_scrambleAcc_avx2(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 31) == 0); + { __m256i* const xacc = (__m256i*) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ + const __m256i* const xsecret = (const __m256i *) secret; + const __m256i prime32 = _mm256_set1_epi32((int)XXH_PRIME32_1); + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m256i); i++) { + /* xacc[i] ^= (xacc[i] >> 47) */ + __m256i const acc_vec = xacc[i]; + __m256i const shifted = _mm256_srli_epi64 (acc_vec, 47); + __m256i const data_vec = _mm256_xor_si256 (acc_vec, shifted); + /* xacc[i] ^= xsecret; */ + __m256i const key_vec = _mm256_loadu_si256 (xsecret+i); + __m256i const data_key = _mm256_xor_si256 (data_vec, key_vec); + + /* xacc[i] *= XXH_PRIME32_1; */ + __m256i const data_key_hi = _mm256_srli_epi64 (data_key, 32); + __m256i const prod_lo = _mm256_mul_epu32 (data_key, prime32); + __m256i const prod_hi = _mm256_mul_epu32 (data_key_hi, prime32); + xacc[i] = _mm256_add_epi64(prod_lo, _mm256_slli_epi64(prod_hi, 32)); + } + } +} + +XXH_FORCE_INLINE XXH_TARGET_AVX2 void XXH3_initCustomSecret_avx2(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 31) == 0); + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE / sizeof(__m256i)) == 6); + XXH_STATIC_ASSERT(XXH_SEC_ALIGN <= 64); + (void)(&XXH_writeLE64); + XXH_PREFETCH(customSecret); + { __m256i const seed = _mm256_set_epi64x((xxh_i64)(0U - seed64), (xxh_i64)seed64, (xxh_i64)(0U - seed64), (xxh_i64)seed64); + + const __m256i* const src = (const __m256i*) ((const void*) XXH3_kSecret); + __m256i* dest = ( __m256i*) customSecret; + +# if defined(__GNUC__) || defined(__clang__) + /* + * On GCC & Clang, marking 'dest' as modified will cause the compiler: + * - do not extract the secret from sse registers in the internal loop + * - use less common registers, and avoid pushing these reg into stack + */ + XXH_COMPILER_GUARD(dest); +# endif + XXH_ASSERT(((size_t)src & 31) == 0); /* control alignment */ + XXH_ASSERT(((size_t)dest & 31) == 0); + + /* GCC -O2 need unroll loop manually */ + dest[0] = _mm256_add_epi64(_mm256_load_si256(src+0), seed); + dest[1] = _mm256_add_epi64(_mm256_load_si256(src+1), seed); + dest[2] = _mm256_add_epi64(_mm256_load_si256(src+2), seed); + dest[3] = _mm256_add_epi64(_mm256_load_si256(src+3), seed); + dest[4] = _mm256_add_epi64(_mm256_load_si256(src+4), seed); + dest[5] = _mm256_add_epi64(_mm256_load_si256(src+5), seed); + } +} + +#endif + +/* x86dispatch always generates SSE2 */ +#if (XXH_VECTOR == XXH_SSE2) || defined(XXH_X86DISPATCH) + +#ifndef XXH_TARGET_SSE2 +# define XXH_TARGET_SSE2 /* disable attribute target */ +#endif + +XXH_FORCE_INLINE XXH_TARGET_SSE2 void +XXH3_accumulate_512_sse2( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + /* SSE2 is just a half-scale version of the AVX2 version. */ + XXH_ASSERT((((size_t)acc) & 15) == 0); + { __m128i* const xacc = (__m128i *) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ + const __m128i* const xinput = (const __m128i *) input; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ + const __m128i* const xsecret = (const __m128i *) secret; + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m128i); i++) { + /* data_vec = xinput[i]; */ + __m128i const data_vec = _mm_loadu_si128 (xinput+i); + /* key_vec = xsecret[i]; */ + __m128i const key_vec = _mm_loadu_si128 (xsecret+i); + /* data_key = data_vec ^ key_vec; */ + __m128i const data_key = _mm_xor_si128 (data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m128i const data_key_lo = _mm_shuffle_epi32 (data_key, _MM_SHUFFLE(0, 3, 0, 1)); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m128i const product = _mm_mul_epu32 (data_key, data_key_lo); + /* xacc[i] += swap(data_vec); */ + __m128i const data_swap = _mm_shuffle_epi32(data_vec, _MM_SHUFFLE(1,0,3,2)); + __m128i const sum = _mm_add_epi64(xacc[i], data_swap); + /* xacc[i] += product; */ + xacc[i] = _mm_add_epi64(product, sum); + } } +} +XXH_FORCE_INLINE XXH_TARGET_SSE2 XXH3_ACCUMULATE_TEMPLATE(sse2) + +XXH_FORCE_INLINE XXH_TARGET_SSE2 void +XXH3_scrambleAcc_sse2(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + { __m128i* const xacc = (__m128i*) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ + const __m128i* const xsecret = (const __m128i *) secret; + const __m128i prime32 = _mm_set1_epi32((int)XXH_PRIME32_1); + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m128i); i++) { + /* xacc[i] ^= (xacc[i] >> 47) */ + __m128i const acc_vec = xacc[i]; + __m128i const shifted = _mm_srli_epi64 (acc_vec, 47); + __m128i const data_vec = _mm_xor_si128 (acc_vec, shifted); + /* xacc[i] ^= xsecret[i]; */ + __m128i const key_vec = _mm_loadu_si128 (xsecret+i); + __m128i const data_key = _mm_xor_si128 (data_vec, key_vec); + + /* xacc[i] *= XXH_PRIME32_1; */ + __m128i const data_key_hi = _mm_shuffle_epi32 (data_key, _MM_SHUFFLE(0, 3, 0, 1)); + __m128i const prod_lo = _mm_mul_epu32 (data_key, prime32); + __m128i const prod_hi = _mm_mul_epu32 (data_key_hi, prime32); + xacc[i] = _mm_add_epi64(prod_lo, _mm_slli_epi64(prod_hi, 32)); + } + } +} + +XXH_FORCE_INLINE XXH_TARGET_SSE2 void XXH3_initCustomSecret_sse2(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 15) == 0); + (void)(&XXH_writeLE64); + { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / sizeof(__m128i); + +# if defined(_MSC_VER) && defined(_M_IX86) && _MSC_VER < 1900 + /* MSVC 32bit mode does not support _mm_set_epi64x before 2015 */ + XXH_ALIGN(16) const xxh_i64 seed64x2[2] = { (xxh_i64)seed64, (xxh_i64)(0U - seed64) }; + __m128i const seed = _mm_load_si128((__m128i const*)seed64x2); +# else + __m128i const seed = _mm_set_epi64x((xxh_i64)(0U - seed64), (xxh_i64)seed64); +# endif + int i; + + const void* const src16 = XXH3_kSecret; + __m128i* dst16 = (__m128i*) customSecret; +# if defined(__GNUC__) || defined(__clang__) + /* + * On GCC & Clang, marking 'dest' as modified will cause the compiler: + * - do not extract the secret from sse registers in the internal loop + * - use less common registers, and avoid pushing these reg into stack + */ + XXH_COMPILER_GUARD(dst16); +# endif + XXH_ASSERT(((size_t)src16 & 15) == 0); /* control alignment */ + XXH_ASSERT(((size_t)dst16 & 15) == 0); + + for (i=0; i < nbRounds; ++i) { + dst16[i] = _mm_add_epi64(_mm_load_si128((const __m128i *)src16+i), seed); + } } +} + +#endif + +#if (XXH_VECTOR == XXH_NEON) + +/* forward declarations for the scalar routines */ +XXH_FORCE_INLINE void +XXH3_scalarRound(void* XXH_RESTRICT acc, void const* XXH_RESTRICT input, + void const* XXH_RESTRICT secret, size_t lane); + +XXH_FORCE_INLINE void +XXH3_scalarScrambleRound(void* XXH_RESTRICT acc, + void const* XXH_RESTRICT secret, size_t lane); + +/*! + * @internal + * @brief The bulk processing loop for NEON and WASM SIMD128. + * + * The NEON code path is actually partially scalar when running on AArch64. This + * is to optimize the pipelining and can have up to 15% speedup depending on the + * CPU, and it also mitigates some GCC codegen issues. + * + * @see XXH3_NEON_LANES for configuring this and details about this optimization. + * + * NEON's 32-bit to 64-bit long multiply takes a half vector of 32-bit + * integers instead of the other platforms which mask full 64-bit vectors, + * so the setup is more complicated than just shifting right. + * + * Additionally, there is an optimization for 4 lanes at once noted below. + * + * Since, as stated, the most optimal amount of lanes for Cortexes is 6, + * there needs to be *three* versions of the accumulate operation used + * for the remaining 2 lanes. + * + * WASM's SIMD128 uses SIMDe's arm_neon.h polyfill because the intrinsics overlap + * nearly perfectly. + */ + +XXH_FORCE_INLINE void +XXH3_accumulate_512_neon( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + XXH_STATIC_ASSERT(XXH3_NEON_LANES > 0 && XXH3_NEON_LANES <= XXH_ACC_NB && XXH3_NEON_LANES % 2 == 0); + { /* GCC for darwin arm64 does not like aliasing here */ + xxh_aliasing_uint64x2_t* const xacc = (xxh_aliasing_uint64x2_t*) acc; + /* We don't use a uint32x4_t pointer because it causes bus errors on ARMv7. */ + uint8_t const* xinput = (const uint8_t *) input; + uint8_t const* xsecret = (const uint8_t *) secret; + + size_t i; +#ifdef __wasm_simd128__ + /* + * On WASM SIMD128, Clang emits direct address loads when XXH3_kSecret + * is constant propagated, which results in it converting it to this + * inside the loop: + * + * a = v128.load(XXH3_kSecret + 0 + $secret_offset, offset = 0) + * b = v128.load(XXH3_kSecret + 16 + $secret_offset, offset = 0) + * ... + * + * This requires a full 32-bit address immediate (and therefore a 6 byte + * instruction) as well as an add for each offset. + * + * Putting an asm guard prevents it from folding (at the cost of losing + * the alignment hint), and uses the free offset in `v128.load` instead + * of adding secret_offset each time which overall reduces code size by + * about a kilobyte and improves performance. + */ + XXH_COMPILER_GUARD(xsecret); +#endif + /* Scalar lanes use the normal scalarRound routine */ + for (i = XXH3_NEON_LANES; i < XXH_ACC_NB; i++) { + XXH3_scalarRound(acc, input, secret, i); + } + i = 0; + /* 4 NEON lanes at a time. */ + for (; i+1 < XXH3_NEON_LANES / 2; i+=2) { + /* data_vec = xinput[i]; */ + uint64x2_t data_vec_1 = XXH_vld1q_u64(xinput + (i * 16)); + uint64x2_t data_vec_2 = XXH_vld1q_u64(xinput + ((i+1) * 16)); + /* key_vec = xsecret[i]; */ + uint64x2_t key_vec_1 = XXH_vld1q_u64(xsecret + (i * 16)); + uint64x2_t key_vec_2 = XXH_vld1q_u64(xsecret + ((i+1) * 16)); + /* data_swap = swap(data_vec) */ + uint64x2_t data_swap_1 = vextq_u64(data_vec_1, data_vec_1, 1); + uint64x2_t data_swap_2 = vextq_u64(data_vec_2, data_vec_2, 1); + /* data_key = data_vec ^ key_vec; */ + uint64x2_t data_key_1 = veorq_u64(data_vec_1, key_vec_1); + uint64x2_t data_key_2 = veorq_u64(data_vec_2, key_vec_2); + + /* + * If we reinterpret the 64x2 vectors as 32x4 vectors, we can use a + * de-interleave operation for 4 lanes in 1 step with `vuzpq_u32` to + * get one vector with the low 32 bits of each lane, and one vector + * with the high 32 bits of each lane. + * + * The intrinsic returns a double vector because the original ARMv7-a + * instruction modified both arguments in place. AArch64 and SIMD128 emit + * two instructions from this intrinsic. + * + * [ dk11L | dk11H | dk12L | dk12H ] -> [ dk11L | dk12L | dk21L | dk22L ] + * [ dk21L | dk21H | dk22L | dk22H ] -> [ dk11H | dk12H | dk21H | dk22H ] + */ + uint32x4x2_t unzipped = vuzpq_u32( + vreinterpretq_u32_u64(data_key_1), + vreinterpretq_u32_u64(data_key_2) + ); + /* data_key_lo = data_key & 0xFFFFFFFF */ + uint32x4_t data_key_lo = unzipped.val[0]; + /* data_key_hi = data_key >> 32 */ + uint32x4_t data_key_hi = unzipped.val[1]; + /* + * Then, we can split the vectors horizontally and multiply which, as for most + * widening intrinsics, have a variant that works on both high half vectors + * for free on AArch64. A similar instruction is available on SIMD128. + * + * sum = data_swap + (u64x2) data_key_lo * (u64x2) data_key_hi + */ + uint64x2_t sum_1 = XXH_vmlal_low_u32(data_swap_1, data_key_lo, data_key_hi); + uint64x2_t sum_2 = XXH_vmlal_high_u32(data_swap_2, data_key_lo, data_key_hi); + /* + * Clang reorders + * a += b * c; // umlal swap.2d, dkl.2s, dkh.2s + * c += a; // add acc.2d, acc.2d, swap.2d + * to + * c += a; // add acc.2d, acc.2d, swap.2d + * c += b * c; // umlal acc.2d, dkl.2s, dkh.2s + * + * While it would make sense in theory since the addition is faster, + * for reasons likely related to umlal being limited to certain NEON + * pipelines, this is worse. A compiler guard fixes this. + */ + XXH_COMPILER_GUARD_CLANG_NEON(sum_1); + XXH_COMPILER_GUARD_CLANG_NEON(sum_2); + /* xacc[i] = acc_vec + sum; */ + xacc[i] = vaddq_u64(xacc[i], sum_1); + xacc[i+1] = vaddq_u64(xacc[i+1], sum_2); + } + /* Operate on the remaining NEON lanes 2 at a time. */ + for (; i < XXH3_NEON_LANES / 2; i++) { + /* data_vec = xinput[i]; */ + uint64x2_t data_vec = XXH_vld1q_u64(xinput + (i * 16)); + /* key_vec = xsecret[i]; */ + uint64x2_t key_vec = XXH_vld1q_u64(xsecret + (i * 16)); + /* acc_vec_2 = swap(data_vec) */ + uint64x2_t data_swap = vextq_u64(data_vec, data_vec, 1); + /* data_key = data_vec ^ key_vec; */ + uint64x2_t data_key = veorq_u64(data_vec, key_vec); + /* For two lanes, just use VMOVN and VSHRN. */ + /* data_key_lo = data_key & 0xFFFFFFFF; */ + uint32x2_t data_key_lo = vmovn_u64(data_key); + /* data_key_hi = data_key >> 32; */ + uint32x2_t data_key_hi = vshrn_n_u64(data_key, 32); + /* sum = data_swap + (u64x2) data_key_lo * (u64x2) data_key_hi; */ + uint64x2_t sum = vmlal_u32(data_swap, data_key_lo, data_key_hi); + /* Same Clang workaround as before */ + XXH_COMPILER_GUARD_CLANG_NEON(sum); + /* xacc[i] = acc_vec + sum; */ + xacc[i] = vaddq_u64 (xacc[i], sum); + } + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(neon) + +XXH_FORCE_INLINE void +XXH3_scrambleAcc_neon(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + + { xxh_aliasing_uint64x2_t* xacc = (xxh_aliasing_uint64x2_t*) acc; + uint8_t const* xsecret = (uint8_t const*) secret; + + size_t i; + /* WASM uses operator overloads and doesn't need these. */ +#ifndef __wasm_simd128__ + /* { prime32_1, prime32_1 } */ + uint32x2_t const kPrimeLo = vdup_n_u32(XXH_PRIME32_1); + /* { 0, prime32_1, 0, prime32_1 } */ + uint32x4_t const kPrimeHi = vreinterpretq_u32_u64(vdupq_n_u64((xxh_u64)XXH_PRIME32_1 << 32)); +#endif + + /* AArch64 uses both scalar and neon at the same time */ + for (i = XXH3_NEON_LANES; i < XXH_ACC_NB; i++) { + XXH3_scalarScrambleRound(acc, secret, i); + } + for (i=0; i < XXH3_NEON_LANES / 2; i++) { + /* xacc[i] ^= (xacc[i] >> 47); */ + uint64x2_t acc_vec = xacc[i]; + uint64x2_t shifted = vshrq_n_u64(acc_vec, 47); + uint64x2_t data_vec = veorq_u64(acc_vec, shifted); + + /* xacc[i] ^= xsecret[i]; */ + uint64x2_t key_vec = XXH_vld1q_u64(xsecret + (i * 16)); + uint64x2_t data_key = veorq_u64(data_vec, key_vec); + /* xacc[i] *= XXH_PRIME32_1 */ +#ifdef __wasm_simd128__ + /* SIMD128 has multiply by u64x2, use it instead of expanding and scalarizing */ + xacc[i] = data_key * XXH_PRIME32_1; +#else + /* + * Expanded version with portable NEON intrinsics + * + * lo(x) * lo(y) + (hi(x) * lo(y) << 32) + * + * prod_hi = hi(data_key) * lo(prime) << 32 + * + * Since we only need 32 bits of this multiply a trick can be used, reinterpreting the vector + * as a uint32x4_t and multiplying by { 0, prime, 0, prime } to cancel out the unwanted bits + * and avoid the shift. + */ + uint32x4_t prod_hi = vmulq_u32 (vreinterpretq_u32_u64(data_key), kPrimeHi); + /* Extract low bits for vmlal_u32 */ + uint32x2_t data_key_lo = vmovn_u64(data_key); + /* xacc[i] = prod_hi + lo(data_key) * XXH_PRIME32_1; */ + xacc[i] = vmlal_u32(vreinterpretq_u64_u32(prod_hi), data_key_lo, kPrimeLo); +#endif + } + } +} +#endif + +#if (XXH_VECTOR == XXH_VSX) + +XXH_FORCE_INLINE void +XXH3_accumulate_512_vsx( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + /* presumed aligned */ + xxh_aliasing_u64x2* const xacc = (xxh_aliasing_u64x2*) acc; + xxh_u8 const* const xinput = (xxh_u8 const*) input; /* no alignment restriction */ + xxh_u8 const* const xsecret = (xxh_u8 const*) secret; /* no alignment restriction */ + xxh_u64x2 const v32 = { 32, 32 }; + size_t i; + for (i = 0; i < XXH_STRIPE_LEN / sizeof(xxh_u64x2); i++) { + /* data_vec = xinput[i]; */ + xxh_u64x2 const data_vec = XXH_vec_loadu(xinput + 16*i); + /* key_vec = xsecret[i]; */ + xxh_u64x2 const key_vec = XXH_vec_loadu(xsecret + 16*i); + xxh_u64x2 const data_key = data_vec ^ key_vec; + /* shuffled = (data_key << 32) | (data_key >> 32); */ + xxh_u32x4 const shuffled = (xxh_u32x4)vec_rl(data_key, v32); + /* product = ((xxh_u64x2)data_key & 0xFFFFFFFF) * ((xxh_u64x2)shuffled & 0xFFFFFFFF); */ + xxh_u64x2 const product = XXH_vec_mulo((xxh_u32x4)data_key, shuffled); + /* acc_vec = xacc[i]; */ + xxh_u64x2 acc_vec = xacc[i]; + acc_vec += product; + + /* swap high and low halves */ +#ifdef __s390x__ + acc_vec += vec_permi(data_vec, data_vec, 2); +#else + acc_vec += vec_xxpermdi(data_vec, data_vec, 2); +#endif + xacc[i] = acc_vec; + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(vsx) + +XXH_FORCE_INLINE void +XXH3_scrambleAcc_vsx(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + + { xxh_aliasing_u64x2* const xacc = (xxh_aliasing_u64x2*) acc; + const xxh_u8* const xsecret = (const xxh_u8*) secret; + /* constants */ + xxh_u64x2 const v32 = { 32, 32 }; + xxh_u64x2 const v47 = { 47, 47 }; + xxh_u32x4 const prime = { XXH_PRIME32_1, XXH_PRIME32_1, XXH_PRIME32_1, XXH_PRIME32_1 }; + size_t i; + for (i = 0; i < XXH_STRIPE_LEN / sizeof(xxh_u64x2); i++) { + /* xacc[i] ^= (xacc[i] >> 47); */ + xxh_u64x2 const acc_vec = xacc[i]; + xxh_u64x2 const data_vec = acc_vec ^ (acc_vec >> v47); + + /* xacc[i] ^= xsecret[i]; */ + xxh_u64x2 const key_vec = XXH_vec_loadu(xsecret + 16*i); + xxh_u64x2 const data_key = data_vec ^ key_vec; + + /* xacc[i] *= XXH_PRIME32_1 */ + /* prod_lo = ((xxh_u64x2)data_key & 0xFFFFFFFF) * ((xxh_u64x2)prime & 0xFFFFFFFF); */ + xxh_u64x2 const prod_even = XXH_vec_mule((xxh_u32x4)data_key, prime); + /* prod_hi = ((xxh_u64x2)data_key >> 32) * ((xxh_u64x2)prime >> 32); */ + xxh_u64x2 const prod_odd = XXH_vec_mulo((xxh_u32x4)data_key, prime); + xacc[i] = prod_odd + (prod_even << v32); + } } +} + +#endif + +#if (XXH_VECTOR == XXH_SVE) + +XXH_FORCE_INLINE void +XXH3_accumulate_512_sve( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + uint64_t *xacc = (uint64_t *)acc; + const uint64_t *xinput = (const uint64_t *)(const void *)input; + const uint64_t *xsecret = (const uint64_t *)(const void *)secret; + svuint64_t kSwap = sveor_n_u64_z(svptrue_b64(), svindex_u64(0, 1), 1); + uint64_t element_count = svcntd(); + if (element_count >= 8) { + svbool_t mask = svptrue_pat_b64(SV_VL8); + svuint64_t vacc = svld1_u64(mask, xacc); + ACCRND(vacc, 0); + svst1_u64(mask, xacc, vacc); + } else if (element_count == 2) { /* sve128 */ + svbool_t mask = svptrue_pat_b64(SV_VL2); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 2); + svuint64_t acc2 = svld1_u64(mask, xacc + 4); + svuint64_t acc3 = svld1_u64(mask, xacc + 6); + ACCRND(acc0, 0); + ACCRND(acc1, 2); + ACCRND(acc2, 4); + ACCRND(acc3, 6); + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 2, acc1); + svst1_u64(mask, xacc + 4, acc2); + svst1_u64(mask, xacc + 6, acc3); + } else { + svbool_t mask = svptrue_pat_b64(SV_VL4); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 4); + ACCRND(acc0, 0); + ACCRND(acc1, 4); + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 4, acc1); + } +} + +XXH_FORCE_INLINE void +XXH3_accumulate_sve(xxh_u64* XXH_RESTRICT acc, + const xxh_u8* XXH_RESTRICT input, + const xxh_u8* XXH_RESTRICT secret, + size_t nbStripes) +{ + if (nbStripes != 0) { + uint64_t *xacc = (uint64_t *)acc; + const uint64_t *xinput = (const uint64_t *)(const void *)input; + const uint64_t *xsecret = (const uint64_t *)(const void *)secret; + svuint64_t kSwap = sveor_n_u64_z(svptrue_b64(), svindex_u64(0, 1), 1); + uint64_t element_count = svcntd(); + if (element_count >= 8) { + svbool_t mask = svptrue_pat_b64(SV_VL8); + svuint64_t vacc = svld1_u64(mask, xacc + 0); + do { + /* svprfd(svbool_t, void *, enum svfprop); */ + svprfd(mask, xinput + 128, SV_PLDL1STRM); + ACCRND(vacc, 0); + xinput += 8; + xsecret += 1; + nbStripes--; + } while (nbStripes != 0); + + svst1_u64(mask, xacc + 0, vacc); + } else if (element_count == 2) { /* sve128 */ + svbool_t mask = svptrue_pat_b64(SV_VL2); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 2); + svuint64_t acc2 = svld1_u64(mask, xacc + 4); + svuint64_t acc3 = svld1_u64(mask, xacc + 6); + do { + svprfd(mask, xinput + 128, SV_PLDL1STRM); + ACCRND(acc0, 0); + ACCRND(acc1, 2); + ACCRND(acc2, 4); + ACCRND(acc3, 6); + xinput += 8; + xsecret += 1; + nbStripes--; + } while (nbStripes != 0); + + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 2, acc1); + svst1_u64(mask, xacc + 4, acc2); + svst1_u64(mask, xacc + 6, acc3); + } else { + svbool_t mask = svptrue_pat_b64(SV_VL4); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 4); + do { + svprfd(mask, xinput + 128, SV_PLDL1STRM); + ACCRND(acc0, 0); + ACCRND(acc1, 4); + xinput += 8; + xsecret += 1; + nbStripes--; + } while (nbStripes != 0); + + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 4, acc1); + } + } +} + +#endif + +/* scalar variants - universal */ + +#if defined(__aarch64__) && (defined(__GNUC__) || defined(__clang__)) +/* + * In XXH3_scalarRound(), GCC and Clang have a similar codegen issue, where they + * emit an excess mask and a full 64-bit multiply-add (MADD X-form). + * + * While this might not seem like much, as AArch64 is a 64-bit architecture, only + * big Cortex designs have a full 64-bit multiplier. + * + * On the little cores, the smaller 32-bit multiplier is used, and full 64-bit + * multiplies expand to 2-3 multiplies in microcode. This has a major penalty + * of up to 4 latency cycles and 2 stall cycles in the multiply pipeline. + * + * Thankfully, AArch64 still provides the 32-bit long multiply-add (UMADDL) which does + * not have this penalty and does the mask automatically. + */ +XXH_FORCE_INLINE xxh_u64 +XXH_mult32to64_add64(xxh_u64 lhs, xxh_u64 rhs, xxh_u64 acc) +{ + xxh_u64 ret; + /* note: %x = 64-bit register, %w = 32-bit register */ + __asm__("umaddl %x0, %w1, %w2, %x3" : "=r" (ret) : "r" (lhs), "r" (rhs), "r" (acc)); + return ret; +} +#else +XXH_FORCE_INLINE xxh_u64 +XXH_mult32to64_add64(xxh_u64 lhs, xxh_u64 rhs, xxh_u64 acc) +{ + return XXH_mult32to64((xxh_u32)lhs, (xxh_u32)rhs) + acc; +} +#endif + +/*! + * @internal + * @brief Scalar round for @ref XXH3_accumulate_512_scalar(). + * + * This is extracted to its own function because the NEON path uses a combination + * of NEON and scalar. + */ +XXH_FORCE_INLINE void +XXH3_scalarRound(void* XXH_RESTRICT acc, + void const* XXH_RESTRICT input, + void const* XXH_RESTRICT secret, + size_t lane) +{ + xxh_u64* xacc = (xxh_u64*) acc; + xxh_u8 const* xinput = (xxh_u8 const*) input; + xxh_u8 const* xsecret = (xxh_u8 const*) secret; + XXH_ASSERT(lane < XXH_ACC_NB); + XXH_ASSERT(((size_t)acc & (XXH_ACC_ALIGN-1)) == 0); + { + xxh_u64 const data_val = XXH_readLE64(xinput + lane * 8); + xxh_u64 const data_key = data_val ^ XXH_readLE64(xsecret + lane * 8); + xacc[lane ^ 1] += data_val; /* swap adjacent lanes */ + xacc[lane] = XXH_mult32to64_add64(data_key /* & 0xFFFFFFFF */, data_key >> 32, xacc[lane]); + } +} + +/*! + * @internal + * @brief Processes a 64 byte block of data using the scalar path. + */ +XXH_FORCE_INLINE void +XXH3_accumulate_512_scalar(void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + size_t i; + /* ARM GCC refuses to unroll this loop, resulting in a 24% slowdown on ARMv6. */ +#if defined(__GNUC__) && !defined(__clang__) \ + && (defined(__arm__) || defined(__thumb2__)) \ + && defined(__ARM_FEATURE_UNALIGNED) /* no unaligned access just wastes bytes */ \ + && XXH_SIZE_OPT <= 0 +# pragma GCC unroll 8 +#endif + for (i=0; i < XXH_ACC_NB; i++) { + XXH3_scalarRound(acc, input, secret, i); + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(scalar) + +/*! + * @internal + * @brief Scalar scramble step for @ref XXH3_scrambleAcc_scalar(). + * + * This is extracted to its own function because the NEON path uses a combination + * of NEON and scalar. + */ +XXH_FORCE_INLINE void +XXH3_scalarScrambleRound(void* XXH_RESTRICT acc, + void const* XXH_RESTRICT secret, + size_t lane) +{ + xxh_u64* const xacc = (xxh_u64*) acc; /* presumed aligned */ + const xxh_u8* const xsecret = (const xxh_u8*) secret; /* no alignment restriction */ + XXH_ASSERT((((size_t)acc) & (XXH_ACC_ALIGN-1)) == 0); + XXH_ASSERT(lane < XXH_ACC_NB); + { + xxh_u64 const key64 = XXH_readLE64(xsecret + lane * 8); + xxh_u64 acc64 = xacc[lane]; + acc64 = XXH_xorshift64(acc64, 47); + acc64 ^= key64; + acc64 *= XXH_PRIME32_1; + xacc[lane] = acc64; + } +} + +/*! + * @internal + * @brief Scrambles the accumulators after a large chunk has been read + */ +XXH_FORCE_INLINE void +XXH3_scrambleAcc_scalar(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + size_t i; + for (i=0; i < XXH_ACC_NB; i++) { + XXH3_scalarScrambleRound(acc, secret, i); + } +} + +XXH_FORCE_INLINE void +XXH3_initCustomSecret_scalar(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + /* + * We need a separate pointer for the hack below, + * which requires a non-const pointer. + * Any decent compiler will optimize this out otherwise. + */ + const xxh_u8* kSecretPtr = XXH3_kSecret; + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 15) == 0); + +#if defined(__GNUC__) && defined(__aarch64__) + /* + * UGLY HACK: + * GCC and Clang generate a bunch of MOV/MOVK pairs for aarch64, and they are + * placed sequentially, in order, at the top of the unrolled loop. + * + * While MOVK is great for generating constants (2 cycles for a 64-bit + * constant compared to 4 cycles for LDR), it fights for bandwidth with + * the arithmetic instructions. + * + * I L S + * MOVK + * MOVK + * MOVK + * MOVK + * ADD + * SUB STR + * STR + * By forcing loads from memory (as the asm line causes the compiler to assume + * that XXH3_kSecretPtr has been changed), the pipelines are used more + * efficiently: + * I L S + * LDR + * ADD LDR + * SUB STR + * STR + * + * See XXH3_NEON_LANES for details on the pipsline. + * + * XXH3_64bits_withSeed, len == 256, Snapdragon 835 + * without hack: 2654.4 MB/s + * with hack: 3202.9 MB/s + */ + XXH_COMPILER_GUARD(kSecretPtr); +#endif + { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / 16; + int i; + for (i=0; i < nbRounds; i++) { + /* + * The asm hack causes the compiler to assume that kSecretPtr aliases with + * customSecret, and on aarch64, this prevented LDP from merging two + * loads together for free. Putting the loads together before the stores + * properly generates LDP. + */ + xxh_u64 lo = XXH_readLE64(kSecretPtr + 16*i) + seed64; + xxh_u64 hi = XXH_readLE64(kSecretPtr + 16*i + 8) - seed64; + XXH_writeLE64((xxh_u8*)customSecret + 16*i, lo); + XXH_writeLE64((xxh_u8*)customSecret + 16*i + 8, hi); + } } +} + + +typedef void (*XXH3_f_accumulate)(xxh_u64* XXH_RESTRICT, const xxh_u8* XXH_RESTRICT, const xxh_u8* XXH_RESTRICT, size_t); +typedef void (*XXH3_f_scrambleAcc)(void* XXH_RESTRICT, const void*); +typedef void (*XXH3_f_initCustomSecret)(void* XXH_RESTRICT, xxh_u64); + + +#if (XXH_VECTOR == XXH_AVX512) + +#define XXH3_accumulate_512 XXH3_accumulate_512_avx512 +#define XXH3_accumulate XXH3_accumulate_avx512 +#define XXH3_scrambleAcc XXH3_scrambleAcc_avx512 +#define XXH3_initCustomSecret XXH3_initCustomSecret_avx512 + +#elif (XXH_VECTOR == XXH_AVX2) + +#define XXH3_accumulate_512 XXH3_accumulate_512_avx2 +#define XXH3_accumulate XXH3_accumulate_avx2 +#define XXH3_scrambleAcc XXH3_scrambleAcc_avx2 +#define XXH3_initCustomSecret XXH3_initCustomSecret_avx2 + +#elif (XXH_VECTOR == XXH_SSE2) + +#define XXH3_accumulate_512 XXH3_accumulate_512_sse2 +#define XXH3_accumulate XXH3_accumulate_sse2 +#define XXH3_scrambleAcc XXH3_scrambleAcc_sse2 +#define XXH3_initCustomSecret XXH3_initCustomSecret_sse2 + +#elif (XXH_VECTOR == XXH_NEON) + +#define XXH3_accumulate_512 XXH3_accumulate_512_neon +#define XXH3_accumulate XXH3_accumulate_neon +#define XXH3_scrambleAcc XXH3_scrambleAcc_neon +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#elif (XXH_VECTOR == XXH_VSX) + +#define XXH3_accumulate_512 XXH3_accumulate_512_vsx +#define XXH3_accumulate XXH3_accumulate_vsx +#define XXH3_scrambleAcc XXH3_scrambleAcc_vsx +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#elif (XXH_VECTOR == XXH_SVE) +#define XXH3_accumulate_512 XXH3_accumulate_512_sve +#define XXH3_accumulate XXH3_accumulate_sve +#define XXH3_scrambleAcc XXH3_scrambleAcc_scalar +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#else /* scalar */ + +#define XXH3_accumulate_512 XXH3_accumulate_512_scalar +#define XXH3_accumulate XXH3_accumulate_scalar +#define XXH3_scrambleAcc XXH3_scrambleAcc_scalar +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#endif + +#if XXH_SIZE_OPT >= 1 /* don't do SIMD for initialization */ +# undef XXH3_initCustomSecret +# define XXH3_initCustomSecret XXH3_initCustomSecret_scalar +#endif + +XXH_FORCE_INLINE void +XXH3_hashLong_internal_loop(xxh_u64* XXH_RESTRICT acc, + const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + size_t const nbStripesPerBlock = (secretSize - XXH_STRIPE_LEN) / XXH_SECRET_CONSUME_RATE; + size_t const block_len = XXH_STRIPE_LEN * nbStripesPerBlock; + size_t const nb_blocks = (len - 1) / block_len; + + size_t n; + + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); + + for (n = 0; n < nb_blocks; n++) { + f_acc(acc, input + n*block_len, secret, nbStripesPerBlock); + f_scramble(acc, secret + secretSize - XXH_STRIPE_LEN); + } + + /* last partial block */ + XXH_ASSERT(len > XXH_STRIPE_LEN); + { size_t const nbStripes = ((len - 1) - (block_len * nb_blocks)) / XXH_STRIPE_LEN; + XXH_ASSERT(nbStripes <= (secretSize / XXH_SECRET_CONSUME_RATE)); + f_acc(acc, input + nb_blocks*block_len, secret, nbStripes); + + /* last stripe */ + { const xxh_u8* const p = input + len - XXH_STRIPE_LEN; +#define XXH_SECRET_LASTACC_START 7 /* not aligned on 8, last secret is different from acc & scrambler */ + XXH3_accumulate_512(acc, p, secret + secretSize - XXH_STRIPE_LEN - XXH_SECRET_LASTACC_START); + } } +} + +XXH_FORCE_INLINE xxh_u64 +XXH3_mix2Accs(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret) +{ + return XXH3_mul128_fold64( + acc[0] ^ XXH_readLE64(secret), + acc[1] ^ XXH_readLE64(secret+8) ); +} + +static XXH64_hash_t +XXH3_mergeAccs(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret, xxh_u64 start) +{ + xxh_u64 result64 = start; + size_t i = 0; + + for (i = 0; i < 4; i++) { + result64 += XXH3_mix2Accs(acc+2*i, secret + 16*i); +#if defined(__clang__) /* Clang */ \ + && (defined(__arm__) || defined(__thumb__)) /* ARMv7 */ \ + && (defined(__ARM_NEON) || defined(__ARM_NEON__)) /* NEON */ \ + && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable */ + /* + * UGLY HACK: + * Prevent autovectorization on Clang ARMv7-a. Exact same problem as + * the one in XXH3_len_129to240_64b. Speeds up shorter keys > 240b. + * XXH3_64bits, len == 256, Snapdragon 835: + * without hack: 2063.7 MB/s + * with hack: 2560.7 MB/s + */ + XXH_COMPILER_GUARD(result64); +#endif + } + + return XXH3_avalanche(result64); +} + +#define XXH3_INIT_ACC { XXH_PRIME32_3, XXH_PRIME64_1, XXH_PRIME64_2, XXH_PRIME64_3, \ + XXH_PRIME64_4, XXH_PRIME32_2, XXH_PRIME64_5, XXH_PRIME32_1 } + +XXH_FORCE_INLINE XXH64_hash_t +XXH3_hashLong_64b_internal(const void* XXH_RESTRICT input, size_t len, + const void* XXH_RESTRICT secret, size_t secretSize, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[XXH_ACC_NB] = XXH3_INIT_ACC; + + XXH3_hashLong_internal_loop(acc, (const xxh_u8*)input, len, (const xxh_u8*)secret, secretSize, f_acc, f_scramble); + + /* converge into final hash */ + XXH_STATIC_ASSERT(sizeof(acc) == 64); + /* do not align on 8, so that the secret is different from the accumulator */ +#define XXH_SECRET_MERGEACCS_START 11 + XXH_ASSERT(secretSize >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); + return XXH3_mergeAccs(acc, (const xxh_u8*)secret + XXH_SECRET_MERGEACCS_START, (xxh_u64)len * XXH_PRIME64_1); +} + +/* + * It's important for performance to transmit secret's size (when it's static) + * so that the compiler can properly optimize the vectorized loop. + * This makes a big performance difference for "medium" keys (<1 KB) when using AVX instruction set. + * When the secret size is unknown, or on GCC 12 where the mix of NO_INLINE and FORCE_INLINE + * breaks -Og, this is XXH_NO_INLINE. + */ +XXH3_WITH_SECRET_INLINE XXH64_hash_t +XXH3_hashLong_64b_withSecret(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; + return XXH3_hashLong_64b_internal(input, len, secret, secretLen, XXH3_accumulate, XXH3_scrambleAcc); +} + +/* + * It's preferable for performance that XXH3_hashLong is not inlined, + * as it results in a smaller function for small data, easier to the instruction cache. + * Note that inside this no_inline function, we do inline the internal loop, + * and provide a statically defined secret size to allow optimization of vector loop. + */ +XXH_NO_INLINE XXH_PUREF XXH64_hash_t +XXH3_hashLong_64b_default(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; (void)secret; (void)secretLen; + return XXH3_hashLong_64b_internal(input, len, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_accumulate, XXH3_scrambleAcc); +} + +/* + * XXH3_hashLong_64b_withSeed(): + * Generate a custom key based on alteration of default XXH3_kSecret with the seed, + * and then use this key for long mode hashing. + * + * This operation is decently fast but nonetheless costs a little bit of time. + * Try to avoid it whenever possible (typically when seed==0). + * + * It's important for performance that XXH3_hashLong is not inlined. Not sure + * why (uop cache maybe?), but the difference is large and easily measurable. + */ +XXH_FORCE_INLINE XXH64_hash_t +XXH3_hashLong_64b_withSeed_internal(const void* input, size_t len, + XXH64_hash_t seed, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble, + XXH3_f_initCustomSecret f_initSec) +{ +#if XXH_SIZE_OPT <= 0 + if (seed == 0) + return XXH3_hashLong_64b_internal(input, len, + XXH3_kSecret, sizeof(XXH3_kSecret), + f_acc, f_scramble); +#endif + { XXH_ALIGN(XXH_SEC_ALIGN) xxh_u8 secret[XXH_SECRET_DEFAULT_SIZE]; + f_initSec(secret, seed); + return XXH3_hashLong_64b_internal(input, len, secret, sizeof(secret), + f_acc, f_scramble); + } +} + +/* + * It's important for performance that XXH3_hashLong is not inlined. + */ +XXH_NO_INLINE XXH64_hash_t +XXH3_hashLong_64b_withSeed(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) +{ + (void)secret; (void)secretLen; + return XXH3_hashLong_64b_withSeed_internal(input, len, seed, + XXH3_accumulate, XXH3_scrambleAcc, XXH3_initCustomSecret); +} + + +typedef XXH64_hash_t (*XXH3_hashLong64_f)(const void* XXH_RESTRICT, size_t, + XXH64_hash_t, const xxh_u8* XXH_RESTRICT, size_t); + +XXH_FORCE_INLINE XXH64_hash_t +XXH3_64bits_internal(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen, + XXH3_hashLong64_f f_hashLong) +{ + XXH_ASSERT(secretLen >= XXH3_SECRET_SIZE_MIN); + /* + * If an action is to be taken if `secretLen` condition is not respected, + * it should be done here. + * For now, it's a contract pre-condition. + * Adding a check and a branch here would cost performance at every hash. + * Also, note that function signature doesn't offer room to return an error. + */ + if (len <= 16) + return XXH3_len_0to16_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, seed64); + if (len <= 128) + return XXH3_len_17to128_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + if (len <= XXH3_MIDSIZE_MAX) + return XXH3_len_129to240_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + return f_hashLong(input, len, seed64, (const xxh_u8*)secret, secretLen); +} + + +/* === Public entry point === */ + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t XXH3_64bits(XXH_NOESCAPE const void* input, size_t length) +{ + return XXH3_64bits_internal(input, length, 0, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_64b_default); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t +XXH3_64bits_withSecret(XXH_NOESCAPE const void* input, size_t length, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + return XXH3_64bits_internal(input, length, 0, secret, secretSize, XXH3_hashLong_64b_withSecret); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t +XXH3_64bits_withSeed(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed) +{ + return XXH3_64bits_internal(input, length, seed, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_64b_withSeed); +} + +XXH_PUBLIC_API XXH64_hash_t +XXH3_64bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t length, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) +{ + if (length <= XXH3_MIDSIZE_MAX) + return XXH3_64bits_internal(input, length, seed, XXH3_kSecret, sizeof(XXH3_kSecret), NULL); + return XXH3_hashLong_64b_withSecret(input, length, seed, (const xxh_u8*)secret, secretSize); +} + + +/* === XXH3 streaming === */ +#ifndef XXH_NO_STREAM +/* + * Malloc's a pointer that is always aligned to align. + * + * This must be freed with `XXH_alignedFree()`. + * + * malloc typically guarantees 16 byte alignment on 64-bit systems and 8 byte + * alignment on 32-bit. This isn't enough for the 32 byte aligned loads in AVX2 + * or on 32-bit, the 16 byte aligned loads in SSE2 and NEON. + * + * This underalignment previously caused a rather obvious crash which went + * completely unnoticed due to XXH3_createState() not actually being tested. + * Credit to RedSpah for noticing this bug. + * + * The alignment is done manually: Functions like posix_memalign or _mm_malloc + * are avoided: To maintain portability, we would have to write a fallback + * like this anyways, and besides, testing for the existence of library + * functions without relying on external build tools is impossible. + * + * The method is simple: Overallocate, manually align, and store the offset + * to the original behind the returned pointer. + * + * Align must be a power of 2 and 8 <= align <= 128. + */ +static XXH_MALLOCF void* XXH_alignedMalloc(size_t s, size_t align) +{ + XXH_ASSERT(align <= 128 && align >= 8); /* range check */ + XXH_ASSERT((align & (align-1)) == 0); /* power of 2 */ + XXH_ASSERT(s != 0 && s < (s + align)); /* empty/overflow */ + { /* Overallocate to make room for manual realignment and an offset byte */ + xxh_u8* base = (xxh_u8*)XXH_malloc(s + align); + if (base != NULL) { + /* + * Get the offset needed to align this pointer. + * + * Even if the returned pointer is aligned, there will always be + * at least one byte to store the offset to the original pointer. + */ + size_t offset = align - ((size_t)base & (align - 1)); /* base % align */ + /* Add the offset for the now-aligned pointer */ + xxh_u8* ptr = base + offset; + + XXH_ASSERT((size_t)ptr % align == 0); + + /* Store the offset immediately before the returned pointer. */ + ptr[-1] = (xxh_u8)offset; + return ptr; + } + return NULL; + } +} +/* + * Frees an aligned pointer allocated by XXH_alignedMalloc(). Don't pass + * normal malloc'd pointers, XXH_alignedMalloc has a specific data layout. + */ +static void XXH_alignedFree(void* p) +{ + if (p != NULL) { + xxh_u8* ptr = (xxh_u8*)p; + /* Get the offset byte we added in XXH_malloc. */ + xxh_u8 offset = ptr[-1]; + /* Free the original malloc'd pointer */ + xxh_u8* base = ptr - offset; + XXH_free(base); + } +} +/*! @ingroup XXH3_family */ +/*! + * @brief Allocate an @ref XXH3_state_t. + * + * @return An allocated pointer of @ref XXH3_state_t on success. + * @return `NULL` on failure. + * + * @note Must be freed with XXH3_freeState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH3_state_t* XXH3_createState(void) +{ + XXH3_state_t* const state = (XXH3_state_t*)XXH_alignedMalloc(sizeof(XXH3_state_t), 64); + if (state==NULL) return NULL; + XXH3_INITSTATE(state); + return state; +} + +/*! @ingroup XXH3_family */ +/*! + * @brief Frees an @ref XXH3_state_t. + * + * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). + * + * @return @ref XXH_OK. + * + * @note Must be allocated with XXH3_createState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr) +{ + XXH_alignedFree(statePtr); + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API void +XXH3_copyState(XXH_NOESCAPE XXH3_state_t* dst_state, XXH_NOESCAPE const XXH3_state_t* src_state) +{ + XXH_memcpy(dst_state, src_state, sizeof(*dst_state)); +} + +static void +XXH3_reset_internal(XXH3_state_t* statePtr, + XXH64_hash_t seed, + const void* secret, size_t secretSize) +{ + size_t const initStart = offsetof(XXH3_state_t, bufferedSize); + size_t const initLength = offsetof(XXH3_state_t, nbStripesPerBlock) - initStart; + XXH_ASSERT(offsetof(XXH3_state_t, nbStripesPerBlock) > initStart); + XXH_ASSERT(statePtr != NULL); + /* set members from bufferedSize to nbStripesPerBlock (excluded) to 0 */ + memset((char*)statePtr + initStart, 0, initLength); + statePtr->acc[0] = XXH_PRIME32_3; + statePtr->acc[1] = XXH_PRIME64_1; + statePtr->acc[2] = XXH_PRIME64_2; + statePtr->acc[3] = XXH_PRIME64_3; + statePtr->acc[4] = XXH_PRIME64_4; + statePtr->acc[5] = XXH_PRIME32_2; + statePtr->acc[6] = XXH_PRIME64_5; + statePtr->acc[7] = XXH_PRIME32_1; + statePtr->seed = seed; + statePtr->useSeed = (seed != 0); + statePtr->extSecret = (const unsigned char*)secret; + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); + statePtr->secretLimit = secretSize - XXH_STRIPE_LEN; + statePtr->nbStripesPerBlock = statePtr->secretLimit / XXH_SECRET_CONSUME_RATE; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr) +{ + if (statePtr == NULL) return XXH_ERROR; + XXH3_reset_internal(statePtr, 0, XXH3_kSecret, XXH_SECRET_DEFAULT_SIZE); + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + if (statePtr == NULL) return XXH_ERROR; + XXH3_reset_internal(statePtr, 0, secret, secretSize); + if (secret == NULL) return XXH_ERROR; + if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed) +{ + if (statePtr == NULL) return XXH_ERROR; + if (seed==0) return XXH3_64bits_reset(statePtr); + if ((seed != statePtr->seed) || (statePtr->extSecret != NULL)) + XXH3_initCustomSecret(statePtr->customSecret, seed); + XXH3_reset_internal(statePtr, seed, NULL, XXH_SECRET_DEFAULT_SIZE); + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed64) +{ + if (statePtr == NULL) return XXH_ERROR; + if (secret == NULL) return XXH_ERROR; + if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; + XXH3_reset_internal(statePtr, seed64, secret, secretSize); + statePtr->useSeed = 1; /* always, even if seed64==0 */ + return XXH_OK; +} + +/*! + * @internal + * @brief Processes a large input for XXH3_update() and XXH3_digest_long(). + * + * Unlike XXH3_hashLong_internal_loop(), this can process data that overlaps a block. + * + * @param acc Pointer to the 8 accumulator lanes + * @param nbStripesSoFarPtr In/out pointer to the number of leftover stripes in the block* + * @param nbStripesPerBlock Number of stripes in a block + * @param input Input pointer + * @param nbStripes Number of stripes to process + * @param secret Secret pointer + * @param secretLimit Offset of the last block in @p secret + * @param f_acc Pointer to an XXH3_accumulate implementation + * @param f_scramble Pointer to an XXH3_scrambleAcc implementation + * @return Pointer past the end of @p input after processing + */ +XXH_FORCE_INLINE const xxh_u8 * +XXH3_consumeStripes(xxh_u64* XXH_RESTRICT acc, + size_t* XXH_RESTRICT nbStripesSoFarPtr, size_t nbStripesPerBlock, + const xxh_u8* XXH_RESTRICT input, size_t nbStripes, + const xxh_u8* XXH_RESTRICT secret, size_t secretLimit, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + const xxh_u8* initialSecret = secret + *nbStripesSoFarPtr * XXH_SECRET_CONSUME_RATE; + /* Process full blocks */ + if (nbStripes >= (nbStripesPerBlock - *nbStripesSoFarPtr)) { + /* Process the initial partial block... */ + size_t nbStripesThisIter = nbStripesPerBlock - *nbStripesSoFarPtr; + + do { + /* Accumulate and scramble */ + f_acc(acc, input, initialSecret, nbStripesThisIter); + f_scramble(acc, secret + secretLimit); + input += nbStripesThisIter * XXH_STRIPE_LEN; + nbStripes -= nbStripesThisIter; + /* Then continue the loop with the full block size */ + nbStripesThisIter = nbStripesPerBlock; + initialSecret = secret; + } while (nbStripes >= nbStripesPerBlock); + *nbStripesSoFarPtr = 0; + } + /* Process a partial block */ + if (nbStripes > 0) { + f_acc(acc, input, initialSecret, nbStripes); + input += nbStripes * XXH_STRIPE_LEN; + *nbStripesSoFarPtr += nbStripes; + } + /* Return end pointer */ + return input; +} + +#ifndef XXH3_STREAM_USE_STACK +# if XXH_SIZE_OPT <= 0 && !defined(__clang__) /* clang doesn't need additional stack space */ +# define XXH3_STREAM_USE_STACK 1 +# endif +#endif +/* + * Both XXH3_64bits_update and XXH3_128bits_update use this routine. + */ +XXH_FORCE_INLINE XXH_errorcode +XXH3_update(XXH3_state_t* XXH_RESTRICT const state, + const xxh_u8* XXH_RESTRICT input, size_t len, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + if (input==NULL) { + XXH_ASSERT(len == 0); + return XXH_OK; + } + + XXH_ASSERT(state != NULL); + { const xxh_u8* const bEnd = input + len; + const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; +#if defined(XXH3_STREAM_USE_STACK) && XXH3_STREAM_USE_STACK >= 1 + /* For some reason, gcc and MSVC seem to suffer greatly + * when operating accumulators directly into state. + * Operating into stack space seems to enable proper optimization. + * clang, on the other hand, doesn't seem to need this trick */ + XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[8]; + XXH_memcpy(acc, state->acc, sizeof(acc)); +#else + xxh_u64* XXH_RESTRICT const acc = state->acc; +#endif + state->totalLen += len; + XXH_ASSERT(state->bufferedSize <= XXH3_INTERNALBUFFER_SIZE); + + /* small input : just fill in tmp buffer */ + if (len <= XXH3_INTERNALBUFFER_SIZE - state->bufferedSize) { + XXH_memcpy(state->buffer + state->bufferedSize, input, len); + state->bufferedSize += (XXH32_hash_t)len; + return XXH_OK; + } + + /* total input is now > XXH3_INTERNALBUFFER_SIZE */ + #define XXH3_INTERNALBUFFER_STRIPES (XXH3_INTERNALBUFFER_SIZE / XXH_STRIPE_LEN) + XXH_STATIC_ASSERT(XXH3_INTERNALBUFFER_SIZE % XXH_STRIPE_LEN == 0); /* clean multiple */ + + /* + * Internal buffer is partially filled (always, except at beginning) + * Complete it, then consume it. + */ + if (state->bufferedSize) { + size_t const loadSize = XXH3_INTERNALBUFFER_SIZE - state->bufferedSize; + XXH_memcpy(state->buffer + state->bufferedSize, input, loadSize); + input += loadSize; + XXH3_consumeStripes(acc, + &state->nbStripesSoFar, state->nbStripesPerBlock, + state->buffer, XXH3_INTERNALBUFFER_STRIPES, + secret, state->secretLimit, + f_acc, f_scramble); + state->bufferedSize = 0; + } + XXH_ASSERT(input < bEnd); + if (bEnd - input > XXH3_INTERNALBUFFER_SIZE) { + size_t nbStripes = (size_t)(bEnd - 1 - input) / XXH_STRIPE_LEN; + input = XXH3_consumeStripes(acc, + &state->nbStripesSoFar, state->nbStripesPerBlock, + input, nbStripes, + secret, state->secretLimit, + f_acc, f_scramble); + XXH_memcpy(state->buffer + sizeof(state->buffer) - XXH_STRIPE_LEN, input - XXH_STRIPE_LEN, XXH_STRIPE_LEN); + + } + /* Some remaining input (always) : buffer it */ + XXH_ASSERT(input < bEnd); + XXH_ASSERT(bEnd - input <= XXH3_INTERNALBUFFER_SIZE); + XXH_ASSERT(state->bufferedSize == 0); + XXH_memcpy(state->buffer, input, (size_t)(bEnd-input)); + state->bufferedSize = (XXH32_hash_t)(bEnd-input); +#if defined(XXH3_STREAM_USE_STACK) && XXH3_STREAM_USE_STACK >= 1 + /* save stack accumulators into state */ + XXH_memcpy(state->acc, acc, sizeof(acc)); +#endif + } + + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_update(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len) +{ + return XXH3_update(state, (const xxh_u8*)input, len, + XXH3_accumulate, XXH3_scrambleAcc); +} + + +XXH_FORCE_INLINE void +XXH3_digest_long (XXH64_hash_t* acc, + const XXH3_state_t* state, + const unsigned char* secret) +{ + xxh_u8 lastStripe[XXH_STRIPE_LEN]; + const xxh_u8* lastStripePtr; + + /* + * Digest on a local copy. This way, the state remains unaltered, and it can + * continue ingesting more input afterwards. + */ + XXH_memcpy(acc, state->acc, sizeof(state->acc)); + if (state->bufferedSize >= XXH_STRIPE_LEN) { + /* Consume remaining stripes then point to remaining data in buffer */ + size_t const nbStripes = (state->bufferedSize - 1) / XXH_STRIPE_LEN; + size_t nbStripesSoFar = state->nbStripesSoFar; + XXH3_consumeStripes(acc, + &nbStripesSoFar, state->nbStripesPerBlock, + state->buffer, nbStripes, + secret, state->secretLimit, + XXH3_accumulate, XXH3_scrambleAcc); + lastStripePtr = state->buffer + state->bufferedSize - XXH_STRIPE_LEN; + } else { /* bufferedSize < XXH_STRIPE_LEN */ + /* Copy to temp buffer */ + size_t const catchupSize = XXH_STRIPE_LEN - state->bufferedSize; + XXH_ASSERT(state->bufferedSize > 0); /* there is always some input buffered */ + XXH_memcpy(lastStripe, state->buffer + sizeof(state->buffer) - catchupSize, catchupSize); + XXH_memcpy(lastStripe + catchupSize, state->buffer, state->bufferedSize); + lastStripePtr = lastStripe; + } + /* Last stripe */ + XXH3_accumulate_512(acc, + lastStripePtr, + secret + state->secretLimit - XXH_SECRET_LASTACC_START); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_digest (XXH_NOESCAPE const XXH3_state_t* state) +{ + const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; + if (state->totalLen > XXH3_MIDSIZE_MAX) { + XXH_ALIGN(XXH_ACC_ALIGN) XXH64_hash_t acc[XXH_ACC_NB]; + XXH3_digest_long(acc, state, secret); + return XXH3_mergeAccs(acc, + secret + XXH_SECRET_MERGEACCS_START, + (xxh_u64)state->totalLen * XXH_PRIME64_1); + } + /* totalLen <= XXH3_MIDSIZE_MAX: digesting a short input */ + if (state->useSeed) + return XXH3_64bits_withSeed(state->buffer, (size_t)state->totalLen, state->seed); + return XXH3_64bits_withSecret(state->buffer, (size_t)(state->totalLen), + secret, state->secretLimit + XXH_STRIPE_LEN); +} +#endif /* !XXH_NO_STREAM */ + + +/* ========================================== + * XXH3 128 bits (a.k.a XXH128) + * ========================================== + * XXH3's 128-bit variant has better mixing and strength than the 64-bit variant, + * even without counting the significantly larger output size. + * + * For example, extra steps are taken to avoid the seed-dependent collisions + * in 17-240 byte inputs (See XXH3_mix16B and XXH128_mix32B). + * + * This strength naturally comes at the cost of some speed, especially on short + * lengths. Note that longer hashes are about as fast as the 64-bit version + * due to it using only a slight modification of the 64-bit loop. + * + * XXH128 is also more oriented towards 64-bit machines. It is still extremely + * fast for a _128-bit_ hash on 32-bit (it usually clears XXH64). + */ + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_1to3_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + /* A doubled version of 1to3_64b with different constants. */ + XXH_ASSERT(input != NULL); + XXH_ASSERT(1 <= len && len <= 3); + XXH_ASSERT(secret != NULL); + /* + * len = 1: combinedl = { input[0], 0x01, input[0], input[0] } + * len = 2: combinedl = { input[1], 0x02, input[0], input[1] } + * len = 3: combinedl = { input[2], 0x03, input[0], input[1] } + */ + { xxh_u8 const c1 = input[0]; + xxh_u8 const c2 = input[len >> 1]; + xxh_u8 const c3 = input[len - 1]; + xxh_u32 const combinedl = ((xxh_u32)c1 <<16) | ((xxh_u32)c2 << 24) + | ((xxh_u32)c3 << 0) | ((xxh_u32)len << 8); + xxh_u32 const combinedh = XXH_rotl32(XXH_swap32(combinedl), 13); + xxh_u64 const bitflipl = (XXH_readLE32(secret) ^ XXH_readLE32(secret+4)) + seed; + xxh_u64 const bitfliph = (XXH_readLE32(secret+8) ^ XXH_readLE32(secret+12)) - seed; + xxh_u64 const keyed_lo = (xxh_u64)combinedl ^ bitflipl; + xxh_u64 const keyed_hi = (xxh_u64)combinedh ^ bitfliph; + XXH128_hash_t h128; + h128.low64 = XXH64_avalanche(keyed_lo); + h128.high64 = XXH64_avalanche(keyed_hi); + return h128; + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_4to8_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(4 <= len && len <= 8); + seed ^= (xxh_u64)XXH_swap32((xxh_u32)seed) << 32; + { xxh_u32 const input_lo = XXH_readLE32(input); + xxh_u32 const input_hi = XXH_readLE32(input + len - 4); + xxh_u64 const input_64 = input_lo + ((xxh_u64)input_hi << 32); + xxh_u64 const bitflip = (XXH_readLE64(secret+16) ^ XXH_readLE64(secret+24)) + seed; + xxh_u64 const keyed = input_64 ^ bitflip; + + /* Shift len to the left to ensure it is even, this avoids even multiplies. */ + XXH128_hash_t m128 = XXH_mult64to128(keyed, XXH_PRIME64_1 + (len << 2)); + + m128.high64 += (m128.low64 << 1); + m128.low64 ^= (m128.high64 >> 3); + + m128.low64 = XXH_xorshift64(m128.low64, 35); + m128.low64 *= PRIME_MX2; + m128.low64 = XXH_xorshift64(m128.low64, 28); + m128.high64 = XXH3_avalanche(m128.high64); + return m128; + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_9to16_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(9 <= len && len <= 16); + { xxh_u64 const bitflipl = (XXH_readLE64(secret+32) ^ XXH_readLE64(secret+40)) - seed; + xxh_u64 const bitfliph = (XXH_readLE64(secret+48) ^ XXH_readLE64(secret+56)) + seed; + xxh_u64 const input_lo = XXH_readLE64(input); + xxh_u64 input_hi = XXH_readLE64(input + len - 8); + XXH128_hash_t m128 = XXH_mult64to128(input_lo ^ input_hi ^ bitflipl, XXH_PRIME64_1); + /* + * Put len in the middle of m128 to ensure that the length gets mixed to + * both the low and high bits in the 128x64 multiply below. + */ + m128.low64 += (xxh_u64)(len - 1) << 54; + input_hi ^= bitfliph; + /* + * Add the high 32 bits of input_hi to the high 32 bits of m128, then + * add the long product of the low 32 bits of input_hi and XXH_PRIME32_2 to + * the high 64 bits of m128. + * + * The best approach to this operation is different on 32-bit and 64-bit. + */ + if (sizeof(void *) < sizeof(xxh_u64)) { /* 32-bit */ + /* + * 32-bit optimized version, which is more readable. + * + * On 32-bit, it removes an ADC and delays a dependency between the two + * halves of m128.high64, but it generates an extra mask on 64-bit. + */ + m128.high64 += (input_hi & 0xFFFFFFFF00000000ULL) + XXH_mult32to64((xxh_u32)input_hi, XXH_PRIME32_2); + } else { + /* + * 64-bit optimized (albeit more confusing) version. + * + * Uses some properties of addition and multiplication to remove the mask: + * + * Let: + * a = input_hi.lo = (input_hi & 0x00000000FFFFFFFF) + * b = input_hi.hi = (input_hi & 0xFFFFFFFF00000000) + * c = XXH_PRIME32_2 + * + * a + (b * c) + * Inverse Property: x + y - x == y + * a + (b * (1 + c - 1)) + * Distributive Property: x * (y + z) == (x * y) + (x * z) + * a + (b * 1) + (b * (c - 1)) + * Identity Property: x * 1 == x + * a + b + (b * (c - 1)) + * + * Substitute a, b, and c: + * input_hi.hi + input_hi.lo + ((xxh_u64)input_hi.lo * (XXH_PRIME32_2 - 1)) + * + * Since input_hi.hi + input_hi.lo == input_hi, we get this: + * input_hi + ((xxh_u64)input_hi.lo * (XXH_PRIME32_2 - 1)) + */ + m128.high64 += input_hi + XXH_mult32to64((xxh_u32)input_hi, XXH_PRIME32_2 - 1); + } + /* m128 ^= XXH_swap64(m128 >> 64); */ + m128.low64 ^= XXH_swap64(m128.high64); + + { /* 128x64 multiply: h128 = m128 * XXH_PRIME64_2; */ + XXH128_hash_t h128 = XXH_mult64to128(m128.low64, XXH_PRIME64_2); + h128.high64 += m128.high64 * XXH_PRIME64_2; + + h128.low64 = XXH3_avalanche(h128.low64); + h128.high64 = XXH3_avalanche(h128.high64); + return h128; + } } +} + +/* + * Assumption: `secret` size is >= XXH3_SECRET_SIZE_MIN + */ +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_0to16_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(len <= 16); + { if (len > 8) return XXH3_len_9to16_128b(input, len, secret, seed); + if (len >= 4) return XXH3_len_4to8_128b(input, len, secret, seed); + if (len) return XXH3_len_1to3_128b(input, len, secret, seed); + { XXH128_hash_t h128; + xxh_u64 const bitflipl = XXH_readLE64(secret+64) ^ XXH_readLE64(secret+72); + xxh_u64 const bitfliph = XXH_readLE64(secret+80) ^ XXH_readLE64(secret+88); + h128.low64 = XXH64_avalanche(seed ^ bitflipl); + h128.high64 = XXH64_avalanche( seed ^ bitfliph); + return h128; + } } +} + +/* + * A bit slower than XXH3_mix16B, but handles multiply by zero better. + */ +XXH_FORCE_INLINE XXH128_hash_t +XXH128_mix32B(XXH128_hash_t acc, const xxh_u8* input_1, const xxh_u8* input_2, + const xxh_u8* secret, XXH64_hash_t seed) +{ + acc.low64 += XXH3_mix16B (input_1, secret+0, seed); + acc.low64 ^= XXH_readLE64(input_2) + XXH_readLE64(input_2 + 8); + acc.high64 += XXH3_mix16B (input_2, secret+16, seed); + acc.high64 ^= XXH_readLE64(input_1) + XXH_readLE64(input_1 + 8); + return acc; +} + + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_17to128_128b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(16 < len && len <= 128); + + { XXH128_hash_t acc; + acc.low64 = len * XXH_PRIME64_1; + acc.high64 = 0; + +#if XXH_SIZE_OPT >= 1 + { + /* Smaller, but slightly slower. */ + unsigned int i = (unsigned int)(len - 1) / 32; + do { + acc = XXH128_mix32B(acc, input+16*i, input+len-16*(i+1), secret+32*i, seed); + } while (i-- != 0); + } +#else + if (len > 32) { + if (len > 64) { + if (len > 96) { + acc = XXH128_mix32B(acc, input+48, input+len-64, secret+96, seed); + } + acc = XXH128_mix32B(acc, input+32, input+len-48, secret+64, seed); + } + acc = XXH128_mix32B(acc, input+16, input+len-32, secret+32, seed); + } + acc = XXH128_mix32B(acc, input, input+len-16, secret, seed); +#endif + { XXH128_hash_t h128; + h128.low64 = acc.low64 + acc.high64; + h128.high64 = (acc.low64 * XXH_PRIME64_1) + + (acc.high64 * XXH_PRIME64_4) + + ((len - seed) * XXH_PRIME64_2); + h128.low64 = XXH3_avalanche(h128.low64); + h128.high64 = (XXH64_hash_t)0 - XXH3_avalanche(h128.high64); + return h128; + } + } +} + +XXH_NO_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_129to240_128b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); + + { XXH128_hash_t acc; + unsigned i; + acc.low64 = len * XXH_PRIME64_1; + acc.high64 = 0; + /* + * We set as `i` as offset + 32. We do this so that unchanged + * `len` can be used as upper bound. This reaches a sweet spot + * where both x86 and aarch64 get simple agen and good codegen + * for the loop. + */ + for (i = 32; i < 160; i += 32) { + acc = XXH128_mix32B(acc, + input + i - 32, + input + i - 16, + secret + i - 32, + seed); + } + acc.low64 = XXH3_avalanche(acc.low64); + acc.high64 = XXH3_avalanche(acc.high64); + /* + * NB: `i <= len` will duplicate the last 32-bytes if + * len % 32 was zero. This is an unfortunate necessity to keep + * the hash result stable. + */ + for (i=160; i <= len; i += 32) { + acc = XXH128_mix32B(acc, + input + i - 32, + input + i - 16, + secret + XXH3_MIDSIZE_STARTOFFSET + i - 160, + seed); + } + /* last bytes */ + acc = XXH128_mix32B(acc, + input + len - 16, + input + len - 32, + secret + XXH3_SECRET_SIZE_MIN - XXH3_MIDSIZE_LASTOFFSET - 16, + (XXH64_hash_t)0 - seed); + + { XXH128_hash_t h128; + h128.low64 = acc.low64 + acc.high64; + h128.high64 = (acc.low64 * XXH_PRIME64_1) + + (acc.high64 * XXH_PRIME64_4) + + ((len - seed) * XXH_PRIME64_2); + h128.low64 = XXH3_avalanche(h128.low64); + h128.high64 = (XXH64_hash_t)0 - XXH3_avalanche(h128.high64); + return h128; + } + } +} + +XXH_FORCE_INLINE XXH128_hash_t +XXH3_hashLong_128b_internal(const void* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[XXH_ACC_NB] = XXH3_INIT_ACC; + + XXH3_hashLong_internal_loop(acc, (const xxh_u8*)input, len, secret, secretSize, f_acc, f_scramble); + + /* converge into final hash */ + XXH_STATIC_ASSERT(sizeof(acc) == 64); + XXH_ASSERT(secretSize >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); + { XXH128_hash_t h128; + h128.low64 = XXH3_mergeAccs(acc, + secret + XXH_SECRET_MERGEACCS_START, + (xxh_u64)len * XXH_PRIME64_1); + h128.high64 = XXH3_mergeAccs(acc, + secret + secretSize + - sizeof(acc) - XXH_SECRET_MERGEACCS_START, + ~((xxh_u64)len * XXH_PRIME64_2)); + return h128; + } +} + +/* + * It's important for performance that XXH3_hashLong() is not inlined. + */ +XXH_NO_INLINE XXH_PUREF XXH128_hash_t +XXH3_hashLong_128b_default(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, + const void* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; (void)secret; (void)secretLen; + return XXH3_hashLong_128b_internal(input, len, XXH3_kSecret, sizeof(XXH3_kSecret), + XXH3_accumulate, XXH3_scrambleAcc); +} + +/* + * It's important for performance to pass @p secretLen (when it's static) + * to the compiler, so that it can properly optimize the vectorized loop. + * + * When the secret size is unknown, or on GCC 12 where the mix of NO_INLINE and FORCE_INLINE + * breaks -Og, this is XXH_NO_INLINE. + */ +XXH3_WITH_SECRET_INLINE XXH128_hash_t +XXH3_hashLong_128b_withSecret(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, + const void* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; + return XXH3_hashLong_128b_internal(input, len, (const xxh_u8*)secret, secretLen, + XXH3_accumulate, XXH3_scrambleAcc); +} + +XXH_FORCE_INLINE XXH128_hash_t +XXH3_hashLong_128b_withSeed_internal(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble, + XXH3_f_initCustomSecret f_initSec) +{ + if (seed64 == 0) + return XXH3_hashLong_128b_internal(input, len, + XXH3_kSecret, sizeof(XXH3_kSecret), + f_acc, f_scramble); + { XXH_ALIGN(XXH_SEC_ALIGN) xxh_u8 secret[XXH_SECRET_DEFAULT_SIZE]; + f_initSec(secret, seed64); + return XXH3_hashLong_128b_internal(input, len, (const xxh_u8*)secret, sizeof(secret), + f_acc, f_scramble); + } +} + +/* + * It's important for performance that XXH3_hashLong is not inlined. + */ +XXH_NO_INLINE XXH128_hash_t +XXH3_hashLong_128b_withSeed(const void* input, size_t len, + XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen) +{ + (void)secret; (void)secretLen; + return XXH3_hashLong_128b_withSeed_internal(input, len, seed64, + XXH3_accumulate, XXH3_scrambleAcc, XXH3_initCustomSecret); +} + +typedef XXH128_hash_t (*XXH3_hashLong128_f)(const void* XXH_RESTRICT, size_t, + XXH64_hash_t, const void* XXH_RESTRICT, size_t); + +XXH_FORCE_INLINE XXH128_hash_t +XXH3_128bits_internal(const void* input, size_t len, + XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen, + XXH3_hashLong128_f f_hl128) +{ + XXH_ASSERT(secretLen >= XXH3_SECRET_SIZE_MIN); + /* + * If an action is to be taken if `secret` conditions are not respected, + * it should be done here. + * For now, it's a contract pre-condition. + * Adding a check and a branch here would cost performance at every hash. + */ + if (len <= 16) + return XXH3_len_0to16_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, seed64); + if (len <= 128) + return XXH3_len_17to128_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + if (len <= XXH3_MIDSIZE_MAX) + return XXH3_len_129to240_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + return f_hl128(input, len, seed64, secret, secretLen); +} + + +/* === Public XXH128 API === */ + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t XXH3_128bits(XXH_NOESCAPE const void* input, size_t len) +{ + return XXH3_128bits_internal(input, len, 0, + XXH3_kSecret, sizeof(XXH3_kSecret), + XXH3_hashLong_128b_default); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH3_128bits_withSecret(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + return XXH3_128bits_internal(input, len, 0, + (const xxh_u8*)secret, secretSize, + XXH3_hashLong_128b_withSecret); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH3_128bits_withSeed(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) +{ + return XXH3_128bits_internal(input, len, seed, + XXH3_kSecret, sizeof(XXH3_kSecret), + XXH3_hashLong_128b_withSeed); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH3_128bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) +{ + if (len <= XXH3_MIDSIZE_MAX) + return XXH3_128bits_internal(input, len, seed, XXH3_kSecret, sizeof(XXH3_kSecret), NULL); + return XXH3_hashLong_128b_withSecret(input, len, seed, secret, secretSize); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH128(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) +{ + return XXH3_128bits_withSeed(input, len, seed); +} + + +/* === XXH3 128-bit streaming === */ +#ifndef XXH_NO_STREAM +/* + * All initialization and update functions are identical to 64-bit streaming variant. + * The only difference is the finalization routine. + */ + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr) +{ + return XXH3_64bits_reset(statePtr); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + return XXH3_64bits_reset_withSecret(statePtr, secret, secretSize); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed) +{ + return XXH3_64bits_reset_withSeed(statePtr, seed); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) +{ + return XXH3_64bits_reset_withSecretandSeed(statePtr, secret, secretSize, seed); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_update(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len) +{ + return XXH3_64bits_update(state, input, len); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_digest (XXH_NOESCAPE const XXH3_state_t* state) +{ + const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; + if (state->totalLen > XXH3_MIDSIZE_MAX) { + XXH_ALIGN(XXH_ACC_ALIGN) XXH64_hash_t acc[XXH_ACC_NB]; + XXH3_digest_long(acc, state, secret); + XXH_ASSERT(state->secretLimit + XXH_STRIPE_LEN >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); + { XXH128_hash_t h128; + h128.low64 = XXH3_mergeAccs(acc, + secret + XXH_SECRET_MERGEACCS_START, + (xxh_u64)state->totalLen * XXH_PRIME64_1); + h128.high64 = XXH3_mergeAccs(acc, + secret + state->secretLimit + XXH_STRIPE_LEN + - sizeof(acc) - XXH_SECRET_MERGEACCS_START, + ~((xxh_u64)state->totalLen * XXH_PRIME64_2)); + return h128; + } + } + /* len <= XXH3_MIDSIZE_MAX : short code */ + if (state->useSeed) + return XXH3_128bits_withSeed(state->buffer, (size_t)state->totalLen, state->seed); + return XXH3_128bits_withSecret(state->buffer, (size_t)(state->totalLen), + secret, state->secretLimit + XXH_STRIPE_LEN); +} +#endif /* !XXH_NO_STREAM */ +/* 128-bit utility functions */ + +#include /* memcmp, memcpy */ + +/* return : 1 is equal, 0 if different */ +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API int XXH128_isEqual(XXH128_hash_t h1, XXH128_hash_t h2) +{ + /* note : XXH128_hash_t is compact, it has no padding byte */ + return !(memcmp(&h1, &h2, sizeof(h1))); +} + +/* This prototype is compatible with stdlib's qsort(). + * @return : >0 if *h128_1 > *h128_2 + * <0 if *h128_1 < *h128_2 + * =0 if *h128_1 == *h128_2 */ +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API int XXH128_cmp(XXH_NOESCAPE const void* h128_1, XXH_NOESCAPE const void* h128_2) +{ + XXH128_hash_t const h1 = *(const XXH128_hash_t*)h128_1; + XXH128_hash_t const h2 = *(const XXH128_hash_t*)h128_2; + int const hcmp = (h1.high64 > h2.high64) - (h2.high64 > h1.high64); + /* note : bets that, in most cases, hash values are different */ + if (hcmp) return hcmp; + return (h1.low64 > h2.low64) - (h2.low64 > h1.low64); +} + + +/*====== Canonical representation ======*/ +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API void +XXH128_canonicalFromHash(XXH_NOESCAPE XXH128_canonical_t* dst, XXH128_hash_t hash) +{ + XXH_STATIC_ASSERT(sizeof(XXH128_canonical_t) == sizeof(XXH128_hash_t)); + if (XXH_CPU_LITTLE_ENDIAN) { + hash.high64 = XXH_swap64(hash.high64); + hash.low64 = XXH_swap64(hash.low64); + } + XXH_memcpy(dst, &hash.high64, sizeof(hash.high64)); + XXH_memcpy((char*)dst + sizeof(hash.high64), &hash.low64, sizeof(hash.low64)); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH128_hashFromCanonical(XXH_NOESCAPE const XXH128_canonical_t* src) +{ + XXH128_hash_t h; + h.high64 = XXH_readBE64(src); + h.low64 = XXH_readBE64(src->digest + 8); + return h; +} + + + +/* ========================================== + * Secret generators + * ========================================== + */ +#define XXH_MIN(x, y) (((x) > (y)) ? (y) : (x)) + +XXH_FORCE_INLINE void XXH3_combine16(void* dst, XXH128_hash_t h128) +{ + XXH_writeLE64( dst, XXH_readLE64(dst) ^ h128.low64 ); + XXH_writeLE64( (char*)dst+8, XXH_readLE64((char*)dst+8) ^ h128.high64 ); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_generateSecret(XXH_NOESCAPE void* secretBuffer, size_t secretSize, XXH_NOESCAPE const void* customSeed, size_t customSeedSize) +{ +#if (XXH_DEBUGLEVEL >= 1) + XXH_ASSERT(secretBuffer != NULL); + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); +#else + /* production mode, assert() are disabled */ + if (secretBuffer == NULL) return XXH_ERROR; + if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; +#endif + + if (customSeedSize == 0) { + customSeed = XXH3_kSecret; + customSeedSize = XXH_SECRET_DEFAULT_SIZE; + } +#if (XXH_DEBUGLEVEL >= 1) + XXH_ASSERT(customSeed != NULL); +#else + if (customSeed == NULL) return XXH_ERROR; +#endif + + /* Fill secretBuffer with a copy of customSeed - repeat as needed */ + { size_t pos = 0; + while (pos < secretSize) { + size_t const toCopy = XXH_MIN((secretSize - pos), customSeedSize); + memcpy((char*)secretBuffer + pos, customSeed, toCopy); + pos += toCopy; + } } + + { size_t const nbSeg16 = secretSize / 16; + size_t n; + XXH128_canonical_t scrambler; + XXH128_canonicalFromHash(&scrambler, XXH128(customSeed, customSeedSize, 0)); + for (n=0; n /* abort() */ +#include +#include +#include +#include +#include +#include + +#include +#include + +#ifdef __cplusplus +extern "C" { +#endif + +#include "xxhash/xxhash.h" +#include "sha1/sha1.h" +#include "sha256/sha256.h" + +#ifdef __cplusplus +} +#endif + + +// uuid.uuid5(uuid.NAMESPACE_URL, 'en.wikipedia.org/wiki/Llama.cpp') +#define UUID_NAMESPACE_LLAMA_CPP "ef001206-dadc-5f6d-a15f-3359e577d4e5" +#define UUID_NAMESPACE_LLAMA_CPP_HEX 0xef, 0x00, 0x12, 0x06, 0xda, 0xdc, 0x5f, 0x6d, 0xa1, 0x5f, 0x33, 0x59, 0xe5, 0x77, 0xd4, 0xe5 + + +#define HASH_TYPE_SHA256_STR "sha256" +#define HASH_TYPE_SHA1_STR "sha1" +#define HASH_TYPE_XXH64_STR "xxh64" +#define HASH_TYPE_UUID_STR "uuid" + + +typedef enum { + HASH_EXIT_SUCCESS = 0, // All hash has been generated or validated + HASH_EXIT_FAILURE = 1, // Generic Failure + HASH_EXIT_MISMATCH = 2, // Hash mismatched during validation + HASH_EXIT_MANIFEST_MISSING_ENTRY = 3, // Hash attempted validation but missing entry in manifest + HASH_EXIT_MANIFEST_UNKNOWN_HASH = 4, // Manifest is present, but we do not know any hash format within it + HASH_EXIT_MANIFEST_FILE_ERROR = 5 // Manifest is either missing or not a known format +} hash_exit_code_t; + + +typedef enum { + HASH_MANIFEST_NOT_FOUND, + HASH_MANIFEST_MISMATCH, + HASH_MANIFEST_OK, +} hash_manifest_result_t; + + +struct hash_params { + std::string input; + bool xxh64 = false; + bool sha1 = false; + bool sha256 = false; + bool uuid = false; + + bool no_layer = false; + + bool manifest_is_usable = false; + std::string manifest_file; +}; + +struct manifest_check_params { + bool xxh64 = false; + bool sha1 = false; + bool sha256 = false; + bool uuid = false; +}; + +static char const * hash_manifest_result_to_str(hash_manifest_result_t value) { + switch (value) { + case HASH_MANIFEST_NOT_FOUND: return "Not Found"; + case HASH_MANIFEST_MISMATCH: return "Mismatch"; + case HASH_MANIFEST_OK: return "Ok"; + } + return "?"; +} + +static char const * hash_exit_code_to_str(hash_exit_code_t value) { + switch (value) { + case HASH_EXIT_SUCCESS: return "Success"; + case HASH_EXIT_FAILURE: return "Failure"; + case HASH_EXIT_MISMATCH: return "Mismatch"; + case HASH_EXIT_MANIFEST_MISSING_ENTRY: return "Manifest Missing Entry"; + case HASH_EXIT_MANIFEST_UNKNOWN_HASH: return "Manifest Unknown Hash"; + case HASH_EXIT_MANIFEST_FILE_ERROR: return "Manifest File Error"; + } + return "?"; +} + +static void hash_print_usage(const char * executable) { + const hash_params default_params; + printf("\n"); + printf("usage: %s [options] GGUF_IN\n", executable); + printf("\n"); + printf("Hash a GGUF file"); + printf("\n"); + printf("options:\n"); + printf(" -h, --help show this help message and exit\n"); + printf(" --xxh64 use xxh64 hash\n"); + printf(" --sha1 use sha1 hash\n"); + printf(" --sha256 use sha256 hash\n"); + printf(" --all use all hash\n"); + printf(" --no-layer exclude per layer hash\n"); + printf(" --uuid generate UUIDv5 ID\n"); + printf(" -c, --check verify against a manifest\n"); + printf("\n"); +} + +static void hash_params_parse_ex(int argc, const char ** argv, hash_params & params) { + std::string arg; + bool invalid_param = false; + const std::string arg_prefix = "--"; + + int arg_idx = 1; + for (; arg_idx < argc && strncmp(argv[arg_idx], "--", 2) == 0; arg_idx++) { + arg = argv[arg_idx]; + if (arg.compare(0, arg_prefix.size(), arg_prefix) == 0) { + std::replace(arg.begin(), arg.end(), '_', '-'); + } + + bool arg_found = false; + if (arg == "-h" || arg == "--help") { + hash_print_usage(argv[0]); + exit(0); + } + + if (arg == "--xxh64") { + arg_found = true; + params.xxh64 = true; + } + + if (arg == "--sha1") { + arg_found = true; + params.sha1 = true; + } + + if (arg == "--uuid") { + arg_found = true; + params.uuid = true; + } + + if (arg == "--sha256") { + arg_found = true; + params.sha256 = true; + } + + if (arg == "--all") { + arg_found = true; + params.sha256 = true; + params.sha1 = true; + params.xxh64 = true; + } + + if (arg == "--no-layer") { + arg_found = true; + params.no_layer = true; + } + + if (arg == "-c" || arg == "--check") { + if (++arg_idx >= argc) { + invalid_param = true; + break; + } + arg_found = true; + params.manifest_file = argv[arg_idx]; + } + + if (!arg_found) { + throw std::invalid_argument("error: unknown argument: " + arg); + } + } + + if (invalid_param) { + throw std::invalid_argument("error: invalid parameter for argument:" + arg); + } + + if (argc - arg_idx < 1) { + throw std::invalid_argument("error: bad arguments"); + } + + params.input = argv[arg_idx++]; +} + +static bool hash_params_parse(int argc, const char ** argv, hash_params & params) { + bool result = true; + try { + hash_params_parse_ex(argc, argv, params); + } + catch (const std::invalid_argument & ex) { + fprintf(stderr, "%s\n", ex.what()); + hash_print_usage(argv[0]); + exit(EXIT_FAILURE); + } + return result; +} + +static bool manifest_type(const std::string & manifest_file, manifest_check_params & manifest_check) { + if (manifest_file.empty()) { + return false; + } + + std::ifstream file(manifest_file); + if (!file.is_open()) { + return false; + } + + std::string manifest_entry_line; + while (getline(file, manifest_entry_line)) { + // hash_type_str hash_str tensor_name + // e.g. 'xxh64 f66e9cd66a4396a0 test.gguf:tensor_0' + std::istringstream line_stream(manifest_entry_line); + std::string file_hash_type; + if (line_stream >> file_hash_type) { + if (file_hash_type == HASH_TYPE_SHA256_STR) { + manifest_check.sha256 = true; + } else if (file_hash_type == HASH_TYPE_SHA1_STR) { + manifest_check.sha1 = true; + } else if (file_hash_type == HASH_TYPE_XXH64_STR) { + manifest_check.xxh64 = true; + } else if (file_hash_type == HASH_TYPE_UUID_STR) { + manifest_check.uuid = true; + } + } + } + + return true; +} + +static hash_manifest_result_t manifest_verify(const std::string& manifest_file, const std::string& hash_type_str, const std::string& hash_str, const std::string& tensor_name) { + if (manifest_file.empty()) { + return HASH_MANIFEST_NOT_FOUND; + } + + std::ifstream file(manifest_file); + if (!file.is_open()) { + return HASH_MANIFEST_NOT_FOUND; + } + + std::string manifest_entry_line; + while (getline(file, manifest_entry_line)) { + std::istringstream line_stream(manifest_entry_line); + std::string file_hash_type; + std::string file_hash; + std::string file_tensor_name; + if (line_stream >> file_hash_type >> file_hash >> file_tensor_name) { + // Line parsed. Check hash validity + + if (file_hash_type != hash_type_str) { + continue; + } + + if (file_tensor_name != tensor_name) { + continue; + } + + return (file_hash == hash_str) ? HASH_MANIFEST_OK : HASH_MANIFEST_MISMATCH; + } + } + + return HASH_MANIFEST_NOT_FOUND; +} + +static void generate_uuidv5(const unsigned char sha1_digest[20], unsigned char uuid[16]) { + // Ref: https://www.rfc-editor.org/rfc/rfc9562.html#section-5.5 + // Assumes that digest was processed correctly with the expected namespace + for (int i = 0; i < 16; i++) { + uuid[i] = sha1_digest[i]; + } + + // Set bits corresponding to UUID ver 5 + uuid[ 6] &= ~(0xF << 4); + uuid[ 6] |= (5 << 4); + + // Set bits corresponding to UUID variant 0b10XX + uuid[ 8] &= ~(0xc << 4); + uuid[ 8] |= (0x8 << 4); +} + +static hash_exit_code_t gguf_hash(const hash_params & hash_params) { + const std::string & fname = hash_params.input; + struct ggml_context * ctx_data = NULL; + + struct gguf_init_params params = { + /*.no_alloc = */ false, + /*.ctx = */ &ctx_data, + }; + + // xxh64 init + XXH64_state_t* xxh64_model_hash_state = NULL; + if (hash_params.xxh64) { + xxh64_model_hash_state = XXH64_createState(); + if (xxh64_model_hash_state==NULL) { + abort(); + } + + XXH64_hash_t const seed = 0; + if (XXH64_reset(xxh64_model_hash_state, seed) == XXH_ERROR) { + abort(); + } + } + + // sha1 init + SHA1_CTX sha1_model_hash_ctx; + if (hash_params.sha1) { + SHA1Init(&sha1_model_hash_ctx); + } + + // sha256 init + sha256_t sha256_model_hash_ctx; + if (hash_params.sha256) { + sha256_init(&sha256_model_hash_ctx); + } + + // sha1 for uuid init + SHA1_CTX sha1_for_uuid_ctx; + if (hash_params.uuid) { + unsigned char const uuidv5_namespace[] = {UUID_NAMESPACE_LLAMA_CPP_HEX}; + SHA1Init(&sha1_for_uuid_ctx); + SHA1Update( &sha1_for_uuid_ctx, (unsigned char const *)uuidv5_namespace, sizeof(uuidv5_namespace)); + } + + struct gguf_context * ctx = gguf_init_from_file(fname.c_str(), params); + const int n_tensors = gguf_get_n_tensors(ctx); + bool tensor_layer_in_manifest = false; + bool model_in_manifest = false; + bool tensor_layer_has_mismatch = false; + bool model_has_mismatch = false; + for (int i = 0; i < n_tensors; ++i) { + const char * name = gguf_get_tensor_name(ctx, i); + struct ggml_tensor * cur = ggml_get_tensor(ctx_data, name); + auto n_bytes = ggml_nbytes(cur); + auto *raw_data = cur->data; + const std::string tensor_layer_name = fname + ":" + name; + + if (hash_params.xxh64) { + + if (!hash_params.no_layer) { + // Per Layer Hash + XXH64_hash_t hash = XXH64(raw_data, n_bytes, 0); + + char hex_result[17]; + for (int offset = 0; offset < 8; offset++) { + unsigned int shift_bits_by = (8 * (8 - offset - 1)); + snprintf( ( hex_result + (2*offset)), sizeof(hex_result) - (2*offset), "%02x", (unsigned char) (hash >> shift_bits_by)&0xff); + } + + if (hash_params.manifest_is_usable) { + hash_manifest_result_t verify_result = manifest_verify(hash_params.manifest_file, HASH_TYPE_XXH64_STR, hex_result, tensor_layer_name); + + switch (verify_result) { + case HASH_MANIFEST_NOT_FOUND: + break; + case HASH_MANIFEST_MISMATCH: + tensor_layer_in_manifest = true; + tensor_layer_has_mismatch = true; + break; + case HASH_MANIFEST_OK: + tensor_layer_in_manifest = true; + break; + } + + printf("%-8s %-s %s - %s\n", HASH_TYPE_XXH64_STR, hex_result, tensor_layer_name.c_str(), hash_manifest_result_to_str(verify_result)); + } else { + printf("%-8s %-s %s\n", HASH_TYPE_XXH64_STR, hex_result, tensor_layer_name.c_str()); + } + } + + // Overall Model Hash + if (XXH64_update(xxh64_model_hash_state, raw_data, n_bytes) == XXH_ERROR) abort(); + } + + if (hash_params.sha1) { + + if (!hash_params.no_layer) { + // Per Layer Hash + char result[21]; // sha1 outputs 20 bytes + SHA1( result, (const char *)raw_data, n_bytes); + + char hex_result[41] = {0}; + for (int offset = 0; offset < 20; offset++) { + snprintf( ( hex_result + (2*offset)), sizeof(hex_result) - (2*offset), "%02x", result[offset]&0xff); + } + + if (hash_params.manifest_is_usable) { + hash_manifest_result_t verify_result = manifest_verify(hash_params.manifest_file, HASH_TYPE_SHA1_STR, hex_result, tensor_layer_name); + + switch (verify_result) { + case HASH_MANIFEST_NOT_FOUND: + break; + case HASH_MANIFEST_MISMATCH: + tensor_layer_in_manifest = true; + tensor_layer_has_mismatch = true; + break; + case HASH_MANIFEST_OK: + tensor_layer_in_manifest = true; + break; + } + + printf("%-8s %-s %s - %s\n", HASH_TYPE_SHA1_STR, hex_result, tensor_layer_name.c_str(), hash_manifest_result_to_str(verify_result)); + } else { + printf("%-8s %-s %s\n", HASH_TYPE_SHA1_STR, hex_result, tensor_layer_name.c_str()); + } + } + + // Overall Model Hash + SHA1Update( &sha1_model_hash_ctx, (unsigned char const *)raw_data, n_bytes); + } + + if (hash_params.sha256) { + + if (!hash_params.no_layer) { + // Per Layer Hash + unsigned char result[SHA256_DIGEST_SIZE]; // sha256 outputs 32 bytes + sha256_hash((unsigned char*) result, (const unsigned char *)raw_data, n_bytes); + + char hex_result[SHA256_DIGEST_SIZE * 2 + 1] = {0}; + for (int offset = 0; offset < SHA256_DIGEST_SIZE; offset++) { + snprintf( ( hex_result + (2*offset)), sizeof(hex_result) - (2*offset), "%02x", result[offset]&0xff); + } + + if (hash_params.manifest_is_usable) { + hash_manifest_result_t verify_result = manifest_verify(hash_params.manifest_file, HASH_TYPE_SHA256_STR, hex_result, tensor_layer_name); + + switch (verify_result) { + case HASH_MANIFEST_NOT_FOUND: + break; + case HASH_MANIFEST_MISMATCH: + tensor_layer_in_manifest = true; + tensor_layer_has_mismatch = true; + break; + case HASH_MANIFEST_OK: + tensor_layer_in_manifest = true; + break; + } + + printf("%-8s %-s %s - %s\n", HASH_TYPE_SHA256_STR, hex_result, tensor_layer_name.c_str(), hash_manifest_result_to_str(verify_result)); + } else { + printf("%-8s %-s %s\n", HASH_TYPE_SHA256_STR, hex_result, tensor_layer_name.c_str()); + } + } + + // Overall Model Hash + sha256_update( &sha256_model_hash_ctx, (unsigned char const *)raw_data, n_bytes); + } + + if (hash_params.uuid) { + SHA1Update( &sha1_for_uuid_ctx, (unsigned char const *)raw_data, n_bytes); + } + } + + if (hash_params.xxh64) { + XXH64_hash_t const hash = XXH64_digest(xxh64_model_hash_state); + + char hex_result[17]; + for (int offset = 0; offset < 8; offset++) { + unsigned int shift_bits_by = (8 * (8 - offset - 1)); + snprintf( ( hex_result + (2*offset)), sizeof(hex_result) - (2*offset), "%02x", (unsigned char) (hash >> shift_bits_by)&0xff); + } + + if (hash_params.manifest_is_usable) { + hash_manifest_result_t verify_result = manifest_verify(hash_params.manifest_file, HASH_TYPE_XXH64_STR, hex_result, fname); + + switch (verify_result) { + case HASH_MANIFEST_NOT_FOUND: + break; + case HASH_MANIFEST_MISMATCH: + model_in_manifest = true; + model_has_mismatch = true; + break; + case HASH_MANIFEST_OK: + model_in_manifest = true; + break; + } + + printf("%-8s %-s %s - %s\n", HASH_TYPE_XXH64_STR, hex_result, fname.c_str(), hash_manifest_result_to_str(verify_result)); + } else { + printf("%-8s %-s %s\n", HASH_TYPE_XXH64_STR, hex_result, fname.c_str()); + } + } + + if (hash_params.sha1) { + unsigned char result[21]; + SHA1Final(result, &sha1_model_hash_ctx); + + char hex_result[41]; + for (int offset = 0; offset < 20; offset++) { + snprintf( ( hex_result + (2*offset)), sizeof(hex_result) - (2*offset), "%02x", result[offset]&0xff); + } + + if (hash_params.manifest_is_usable) { + hash_manifest_result_t verify_result = manifest_verify(hash_params.manifest_file, HASH_TYPE_SHA1_STR, hex_result, fname); + + switch (verify_result) { + case HASH_MANIFEST_NOT_FOUND: + break; + case HASH_MANIFEST_MISMATCH: + model_in_manifest = true; + model_has_mismatch = true; + break; + case HASH_MANIFEST_OK: + model_in_manifest = true; + break; + } + + printf("%-8s %-s %s - %s\n", HASH_TYPE_SHA1_STR, hex_result, fname.c_str(), hash_manifest_result_to_str(verify_result)); + } else { + printf("%-8s %-s %s\n", HASH_TYPE_SHA1_STR, hex_result, fname.c_str()); + } + } + + if (hash_params.sha256) { + unsigned char result[SHA256_DIGEST_SIZE]; // sha256 outputs 32 bytes + sha256_final( &sha256_model_hash_ctx, result); + + char hex_result[SHA256_DIGEST_SIZE * 2 + 1] = {0}; + for (int offset = 0; offset < SHA256_DIGEST_SIZE; offset++) { + snprintf( ( hex_result + (2*offset)), sizeof(hex_result) - (2*offset), "%02x", result[offset]&0xff); + } + + if (hash_params.manifest_is_usable) { + hash_manifest_result_t verify_result = manifest_verify(hash_params.manifest_file, HASH_TYPE_SHA256_STR, hex_result, fname); + + switch (verify_result) { + case HASH_MANIFEST_NOT_FOUND: + break; + case HASH_MANIFEST_MISMATCH: + model_in_manifest = true; + model_has_mismatch = true; + break; + case HASH_MANIFEST_OK: + model_in_manifest = true; + break; + } + + printf("%-8s %-s %s - %s\n", HASH_TYPE_SHA256_STR, hex_result, fname.c_str(), hash_manifest_result_to_str(verify_result)); + } else { + printf("%-8s %-s %s\n", HASH_TYPE_SHA256_STR, hex_result, fname.c_str()); + } + } + + if (hash_params.uuid) { + unsigned char result[21]; + SHA1Final(result, &sha1_for_uuid_ctx); + + unsigned char uuid[16]; + generate_uuidv5(result, uuid); + + char string_buffer[37] = {0}; + snprintf(string_buffer, sizeof(string_buffer), "%02x%02x%02x%02x-%02x%02x-%02x%02x-%02x%02x-%02x%02x%02x%02x%02x%02x", + uuid[0], uuid[1], uuid[2], uuid[3], + uuid[4], uuid[5], uuid[6], uuid[7], + uuid[8], uuid[9], uuid[10], uuid[11], + uuid[12], uuid[13], uuid[14], uuid[15]); + + if (hash_params.manifest_is_usable) { + hash_manifest_result_t verify_result = manifest_verify(hash_params.manifest_file, HASH_TYPE_SHA256_STR, string_buffer, fname); + + switch (verify_result) { + case HASH_MANIFEST_NOT_FOUND: + break; + case HASH_MANIFEST_MISMATCH: + model_in_manifest = true; + model_has_mismatch = true; + break; + case HASH_MANIFEST_OK: + model_in_manifest = true; + break; + } + + printf("%-8s %-s %s - %s\n", HASH_TYPE_UUID_STR, string_buffer, fname.c_str(), hash_manifest_result_to_str(verify_result)); + } else { + printf("%-8s %-s %s\n", HASH_TYPE_UUID_STR, string_buffer, fname.c_str()); + } + } + + + ggml_free(ctx_data); + gguf_free(ctx); + + + if (hash_params.manifest_is_usable) { + // In hash verification mode + + if (!model_in_manifest) { + // model missing in manifest? + + // Check tensor layer... + if (!tensor_layer_in_manifest) { + // Still missing? Maybe we are reading the wrong manifest. + return HASH_EXIT_MANIFEST_MISSING_ENTRY; + } + + if (tensor_layer_has_mismatch) { + // Per tensor check found error + return HASH_EXIT_FAILURE; + } + + // All per tensor layer checks passed? Sounds good enough. + return HASH_EXIT_SUCCESS; + } + + // Overall model check passed, but let's check per layer just in case + // If missing, we don't care too much as the overall model checked + if (tensor_layer_in_manifest && tensor_layer_has_mismatch) { + return HASH_EXIT_FAILURE; + } + + if (model_has_mismatch) { + // model has failed hash somewhere in the model + return HASH_EXIT_FAILURE; + } + + // All checks appears to be fine + return HASH_EXIT_SUCCESS; + } + + // In hash generation mode + return HASH_EXIT_SUCCESS; +} + +int main(int argc, const char ** argv) { + hash_params params; + manifest_check_params manifest_check; + hash_params_parse(argc, argv, params); + + if (!params.manifest_file.empty()) { + if (!manifest_type(params.manifest_file, manifest_check)) { + printf("ERROR cannot open manifest %s", params.manifest_file.c_str()); + return HASH_EXIT_MANIFEST_FILE_ERROR; + } + + if (!manifest_check.sha256 && !manifest_check.sha1 && !manifest_check.xxh64 && !manifest_check.uuid) { + printf("ERROR manifest does not have any known hash format in %s", params.manifest_file.c_str()); + return HASH_EXIT_MANIFEST_UNKNOWN_HASH; + } + + printf("manifest %s", params.manifest_file.c_str()); + + if (manifest_check.sha256) { + printf(" sha256"); + } + + if (manifest_check.sha1) { + printf(" sha1"); + } + + if (manifest_check.xxh64) { + printf(" xxh64"); + } + + if (manifest_check.uuid) { + printf(" uuid"); + } + + printf("\n"); + + // Autoselect the highest security hash if manifest is provided but + // the user has not specifically defined the hash they care about + if (!params.xxh64 && !params.sha1 && !params.uuid && !params.sha256) { + // User has not selected a specific value, pick most secure hash + if (manifest_check.sha256) { + params.sha256 = true; + } else if (manifest_check.sha1) { + params.sha1 = true; + } else if (manifest_check.xxh64) { + params.xxh64 = true; + } else if (manifest_check.uuid) { + params.uuid = true; + } + } + + params.manifest_is_usable = true; + } + + // By default if no swich argument provided, assume xxh64 + if (!params.xxh64 && !params.sha1 && !params.uuid && !params.sha256) { + params.xxh64 = true; + } + + hash_exit_code_t exit_code = gguf_hash(params); + + if (params.manifest_is_usable) { + printf("\nVerification results for %s - %s\n", params.manifest_file.c_str(), hash_exit_code_to_str(exit_code)); + } + + return exit_code; +} diff --git a/examples/gguf/gguf.cpp b/examples/gguf/gguf.cpp index 575143771..7498f85ef 100644 --- a/examples/gguf/gguf.cpp +++ b/examples/gguf/gguf.cpp @@ -92,6 +92,11 @@ static bool gguf_ex_read_0(const std::string & fname) { struct gguf_context * ctx = gguf_init_from_file(fname.c_str(), params); + if (!ctx) { + fprintf(stderr, "%s: failed to load '%s'\n", __func__, fname.c_str()); + return false; + } + printf("%s: version: %d\n", __func__, gguf_get_version(ctx)); printf("%s: alignment: %zu\n", __func__, gguf_get_alignment(ctx)); printf("%s: data offset: %zu\n", __func__, gguf_get_data_offset(ctx)); diff --git a/examples/imatrix/README.md b/examples/imatrix/README.md index 29602881a..bb5faec94 100644 --- a/examples/imatrix/README.md +++ b/examples/imatrix/README.md @@ -1,6 +1,6 @@ # llama.cpp/examples/imatrix -Compute an importance matrix for a model and given text dataset. Can be used during quantization to enchance the quality of the quantum models. +Compute an importance matrix for a model and given text dataset. Can be used during quantization to enchance the quality of the quantized models. More information is available here: https://github.com/ggerganov/llama.cpp/pull/4861 ## Usage diff --git a/examples/imatrix/imatrix.cpp b/examples/imatrix/imatrix.cpp index 574f5ed9c..6ce1863cf 100644 --- a/examples/imatrix/imatrix.cpp +++ b/examples/imatrix/imatrix.cpp @@ -127,7 +127,7 @@ bool IMatrixCollector::collect_imatrix(struct ggml_tensor * t, bool ask, void * } else if (e.values.size() != (size_t)src1->ne[0]*n_as) { fprintf(stderr, "Oops: inconsistent size for %s (%d vs %d)\n", wname.c_str(), (int)e.values.size(), (int)src1->ne[0]*n_as); - exit(1); //GGML_ASSERT(false); + exit(1); //GGML_ABORT("fatal error"); } if (m_params.verbosity > 1) { printf("%s[%d]: %32s, %s, %5d x %5d, %d\n", __func__, m_last_call, wname.c_str(), ggml_op_name(t->op), (int)src1->ne[0], (int)src1->ne[2], (int)src1->type); @@ -176,7 +176,7 @@ bool IMatrixCollector::collect_imatrix(struct ggml_tensor * t, bool ask, void * } else if (e.values.size() != (size_t)src1->ne[0]) { fprintf(stderr, "Oops: inconsistent size for %s (%d vs %d)\n", wname.c_str(), (int)e.values.size(), (int)src1->ne[0]); - exit(1); //GGML_ASSERT(false); + exit(1); //GGML_ABORT("fatal error"); } ++e.ncall; if (m_params.verbosity > 1) { diff --git a/examples/infill/README.md b/examples/infill/README.md index 74f42d2fc..810a0c5e7 100644 --- a/examples/infill/README.md +++ b/examples/infill/README.md @@ -15,6 +15,7 @@ In this section, we cover the most commonly used options for running the `infill - `-i, --interactive`: Run the program in interactive mode, allowing you to provide input directly and receive real-time responses. - `-n N, --n-predict N`: Set the number of tokens to predict when generating text. Adjusting this value can influence the length of the generated text. - `-c N, --ctx-size N`: Set the size of the prompt context. The default is 512, but LLaMA models were built with a context of 2048, which will provide better results for longer input/inference. +- `--spm-infill`: Use Suffix/Prefix/Middle pattern for infill (instead of Prefix/Suffix/Middle) as some models prefer this. ## Input Prompts diff --git a/examples/infill/infill.cpp b/examples/infill/infill.cpp index 3e82e4a81..dc93d2301 100644 --- a/examples/infill/infill.cpp +++ b/examples/infill/infill.cpp @@ -204,25 +204,23 @@ int main(int argc, char ** argv) { GGML_ASSERT(llama_add_eos_token(model) != 1); LOG("add_bos: %d\n", add_bos); - bool suff_rm_leading_spc = params.escape; - if (suff_rm_leading_spc && params.input_suffix.find_first_of(' ') == 0 && params.input_suffix.size() > 1) { - params.input_suffix.erase(0, 1); - suff_rm_leading_spc = false; - } std::vector embd_inp; + std::vector embd_end; std::vector inp_pfx = ::llama_tokenize(ctx, params.input_prefix, false); std::vector inp_sfx = ::llama_tokenize(ctx, params.input_suffix, false); - const int space_token = 29871; - if (suff_rm_leading_spc && inp_sfx[0] == space_token) { - inp_sfx.erase(inp_sfx.begin()); - } + + GGML_ASSERT(llama_token_prefix(model) >= 0); + GGML_ASSERT(llama_token_suffix(model) >= 0); + inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(model)); - if (add_bos) { - inp_pfx.insert(inp_pfx.begin(), llama_token_bos(model)); - } inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(model)); - embd_inp = inp_pfx; - embd_inp.insert(embd_inp.end(), inp_sfx.begin(), inp_sfx.end()); + + embd_inp = params.spm_infill ? inp_sfx : inp_pfx; + embd_end = params.spm_infill ? inp_pfx : inp_sfx; + if (add_bos) { + embd_inp.insert(embd_inp.begin(), llama_token_bos(model)); + } + embd_inp.insert(embd_inp.end(), embd_end.begin(), embd_end.end()); const llama_token middle_token = llama_token_middle(model); if (middle_token >= 0) { @@ -514,26 +512,21 @@ int main(int argc, char ** argv) { string_process_escapes(params.input_prefix); string_process_escapes(params.input_suffix); } - suff_rm_leading_spc = params.escape; - if (suff_rm_leading_spc && params.input_suffix.find_first_of(' ') == 0 && params.input_suffix.size() > 1) { - params.input_suffix.erase(0, 1); - suff_rm_leading_spc = false; - } + // tokenize new prefix and suffix std::vector inp_pfx = ::llama_tokenize(ctx, params.input_prefix, false); std::vector inp_sfx = ::llama_tokenize(ctx, params.input_suffix, false); - if (suff_rm_leading_spc && inp_sfx[0] == space_token) { - inp_sfx.erase(inp_sfx.begin()); - } - inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(model)); - if (add_bos) { - inp_pfx.insert(inp_pfx.begin(), llama_token_bos(model)); - } - inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(model)); - embd_inp = inp_pfx; - embd_inp.insert(embd_inp.end(), inp_sfx.begin(), inp_sfx.end()); - const llama_token middle_token = llama_token_middle(model); + inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(model)); + inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(model)); + + embd_inp = params.spm_infill ? inp_sfx : inp_pfx; + embd_end = params.spm_infill ? inp_pfx : inp_sfx; + if (add_bos) { + embd_inp.insert(embd_inp.begin(), llama_token_bos(model)); + } + embd_inp.insert(embd_inp.end(), embd_end.begin(), embd_end.end()); + if (middle_token >= 0) { embd_inp.push_back(middle_token); } @@ -657,4 +650,3 @@ int main(int argc, char ** argv) { return 0; } - diff --git a/examples/json-schema-pydantic-example.py b/examples/json_schema_pydantic_example.py similarity index 95% rename from examples/json-schema-pydantic-example.py rename to examples/json_schema_pydantic_example.py index 2a24f8118..19c0bdb5b 100644 --- a/examples/json-schema-pydantic-example.py +++ b/examples/json_schema_pydantic_example.py @@ -1,9 +1,9 @@ # Usage: #! ./llama-server -m some-model.gguf & #! pip install pydantic -#! python json-schema-pydantic-example.py +#! python json_schema_pydantic_example.py -from pydantic import BaseModel, Extra, TypeAdapter +from pydantic import BaseModel, Field, TypeAdapter from annotated_types import MinLen from typing import Annotated, List, Optional import json, requests @@ -17,6 +17,9 @@ if True: The response_model param takes a type (+ supports Pydantic) and behaves just as w/ Instructor (see below) ''' + response_format = None + type_adapter = None + if response_model: type_adapter = TypeAdapter(response_model) schema = type_adapter.json_schema() diff --git a/examples/json_schema_to_grammar.py b/examples/json_schema_to_grammar.py index 92f6e3d47..a8779bf3b 100755 --- a/examples/json_schema_to_grammar.py +++ b/examples/json_schema_to_grammar.py @@ -1,4 +1,6 @@ #!/usr/bin/env python3 +from __future__ import annotations + import argparse import itertools import json @@ -188,7 +190,7 @@ def _generate_min_max_int(min_value: Optional[int], max_value: Optional[int], ou raise RuntimeError("At least one of min_value or max_value must be set") class BuiltinRule: - def __init__(self, content: str, deps: list = None): + def __init__(self, content: str, deps: list | None = None): self.content = content self.deps = deps or [] @@ -231,7 +233,7 @@ GRAMMAR_RANGE_LITERAL_ESCAPE_RE = re.compile(r'[\r\n"\]\-\\]') GRAMMAR_LITERAL_ESCAPES = {'\r': '\\r', '\n': '\\n', '"': '\\"', '-': '\\-', ']': '\\]'} NON_LITERAL_SET = set('|.()[]{}*+?') -ESCAPED_IN_REGEXPS_BUT_NOT_IN_LITERALS = set('[]()|{}*+?') +ESCAPED_IN_REGEXPS_BUT_NOT_IN_LITERALS = set('^$.[]()|{}*+?') class SchemaConverter: @@ -248,7 +250,7 @@ class SchemaConverter: def _format_literal(self, literal): escaped = GRAMMAR_LITERAL_ESCAPE_RE.sub( - lambda m: GRAMMAR_LITERAL_ESCAPES.get(m.group(0)), literal + lambda m: GRAMMAR_LITERAL_ESCAPES.get(m.group(0)) or m.group(0), literal ) return f'"{escaped}"' @@ -403,11 +405,11 @@ class SchemaConverter: i = 0 length = len(pattern) - def to_rule(s: Tuple[str, bool]) -> str: + def to_rule(s: tuple[str, bool]) -> str: (txt, is_literal) = s return "\"" + txt + "\"" if is_literal else txt - def transform() -> Tuple[str, bool]: + def transform() -> tuple[str, bool]: ''' Parse a unit at index i (advancing it), and return its string representation + whether it's a literal. ''' @@ -420,7 +422,7 @@ class SchemaConverter: # We only need a flat structure here to apply repetition operators to the last item, and # to merge literals at the and (we're parsing grouped ( sequences ) recursively and don't treat '|' specially # (GBNF's syntax is luckily very close to regular expressions!) - seq: list[Tuple[str, bool]] = [] + seq: list[tuple[str, bool]] = [] def get_dot(): if self._dotall: @@ -602,7 +604,7 @@ class SchemaConverter: else: add_component(t, is_required=True) - return self._add_rule(rule_name, self._build_object_rule(properties, required, hybrid_name, additional_properties=[])) + return self._add_rule(rule_name, self._build_object_rule(properties, required, hybrid_name, additional_properties=None)) elif schema_type in (None, 'array') and ('items' in schema or 'prefixItems' in schema): items = schema.get('items') or schema['prefixItems'] @@ -691,7 +693,7 @@ class SchemaConverter: required_props = [k for k in sorted_props if k in required] optional_props = [k for k in sorted_props if k not in required] - if additional_properties != False: + if additional_properties is not None and additional_properties != False: sub_name = f'{name}{"-" if name else ""}additional' value_rule = self.visit(additional_properties, f'{sub_name}-value') if isinstance(additional_properties, dict) else \ self._add_primitive('value', PRIMITIVE_RULES['value']) diff --git a/examples/llama-bench/llama-bench.cpp b/examples/llama-bench/llama-bench.cpp index d641a9f12..521fa8880 100644 --- a/examples/llama-bench/llama-bench.cpp +++ b/examples/llama-bench/llama-bench.cpp @@ -23,6 +23,10 @@ #include "ggml-cuda.h" #include "ggml-sycl.h" +#ifdef GGML_USE_CANN +#include "ggml-cann.h" +#endif + // utils static uint64_t get_time_ns() { using clock = std::chrono::high_resolution_clock; @@ -120,6 +124,17 @@ static std::string get_gpu_info() { id += "/"; } } +#endif +#ifdef GGML_USE_CANN + uint32_t count = ggml_backend_cann_get_device_count(); + for (uint32_t i = 0; i < count; i++) { + char buf[128]; + ggml_backend_cann_get_device_description(i, buf, sizeof(buf)); + id += buf; + if (i < count - 1) { + id += "/"; + } + } #endif // TODO: other backends return id; @@ -135,7 +150,7 @@ static const char * output_format_str(output_formats format) { case JSON: return "json"; case MARKDOWN: return "md"; case SQL: return "sql"; - default: GGML_ASSERT(!"invalid output format"); + default: GGML_ABORT("invalid output format"); } } @@ -161,7 +176,7 @@ static const char * split_mode_str(llama_split_mode mode) { case LLAMA_SPLIT_MODE_NONE: return "none"; case LLAMA_SPLIT_MODE_LAYER: return "layer"; case LLAMA_SPLIT_MODE_ROW: return "row"; - default: GGML_ASSERT(!"invalid split mode"); + default: GGML_ABORT("invalid split mode"); } } @@ -1311,7 +1326,7 @@ static std::unique_ptr create_printer(output_formats format) { case SQL: return std::unique_ptr(new sql_printer()); } - GGML_ASSERT(false); + GGML_ABORT("fatal error"); } int main(int argc, char ** argv) { diff --git a/examples/llama.android/llama/CMakeLists.txt b/examples/llama.android/llama/CMakeLists.txt deleted file mode 100644 index a5618cac0..000000000 --- a/examples/llama.android/llama/CMakeLists.txt +++ /dev/null @@ -1,55 +0,0 @@ - -# For more information about using CMake with Android Studio, read the -# documentation: https://d.android.com/studio/projects/add-native-code.html. -# For more examples on how to use CMake, see https://github.com/android/ndk-samples. - -# Sets the minimum CMake version required for this project. -cmake_minimum_required(VERSION 3.22.1) - -# Declares the project name. The project name can be accessed via ${ PROJECT_NAME}, -# Since this is the top level CMakeLists.txt, the project name is also accessible -# with ${CMAKE_PROJECT_NAME} (both CMake variables are in-sync within the top level -# build script scope). -project("llama-android") - -## Fetch latest llama.cpp from GitHub -#include(FetchContent) -#FetchContent_Declare( -# llama -# GIT_REPOSITORY https://github.com/ggerganov/llama.cpp -# GIT_TAG master -#) -# -## Also provides "common" -#FetchContent_MakeAvailable(llama) - -# llama.cpp CI uses the code from the current branch -# ref: https://github.com/ggerganov/llama.cpp/pull/7341#issuecomment-2117617700 -add_subdirectory(../../../../../../ build-llama) - -# Creates and names a library, sets it as either STATIC -# or SHARED, and provides the relative paths to its source code. -# You can define multiple libraries, and CMake builds them for you. -# Gradle automatically packages shared libraries with your APK. -# -# In this top level CMakeLists.txt, ${CMAKE_PROJECT_NAME} is used to define -# the target library name; in the sub-module's CMakeLists.txt, ${PROJECT_NAME} -# is preferred for the same purpose. -# -# In order to load a library into your app from Java/Kotlin, you must call -# System.loadLibrary() and pass the name of the library defined here; -# for GameActivity/NativeActivity derived applications, the same library name must be -# used in the AndroidManifest.xml file. -add_library(${CMAKE_PROJECT_NAME} SHARED - # List C/C++ source files with relative paths to this CMakeLists.txt. - llama-android.cpp) - -# Specifies libraries CMake should link to your target library. You -# can link libraries from various origins, such as libraries defined in this -# build script, prebuilt third-party libraries, or Android system libraries. -target_link_libraries(${CMAKE_PROJECT_NAME} - # List libraries link to the target library - llama - common - android - log) diff --git a/examples/llama.android/llama/src/main/cpp/CMakeLists.txt b/examples/llama.android/llama/src/main/cpp/CMakeLists.txt index 42ebaad49..2de496574 100644 --- a/examples/llama.android/llama/src/main/cpp/CMakeLists.txt +++ b/examples/llama.android/llama/src/main/cpp/CMakeLists.txt @@ -11,15 +11,15 @@ cmake_minimum_required(VERSION 3.22.1) # build script scope). project("llama-android") -include(FetchContent) -FetchContent_Declare( - llama - GIT_REPOSITORY https://github.com/ggerganov/llama.cpp - GIT_TAG master -) +#include(FetchContent) +#FetchContent_Declare( +# llama +# GIT_REPOSITORY https://github.com/ggerganov/llama.cpp +# GIT_TAG master +#) # Also provides "common" -FetchContent_MakeAvailable(llama) +#FetchContent_MakeAvailable(llama) # Creates and names a library, sets it as either STATIC # or SHARED, and provides the relative paths to its source code. @@ -30,6 +30,10 @@ FetchContent_MakeAvailable(llama) # the target library name; in the sub-module's CMakeLists.txt, ${PROJECT_NAME} # is preferred for the same purpose. # + +#load local llama.cpp +add_subdirectory(../../../../../../ build-llama) + # In order to load a library into your app from Java/Kotlin, you must call # System.loadLibrary() and pass the name of the library defined here; # for GameActivity/NativeActivity derived applications, the same library name must be diff --git a/examples/llama.android/llama/src/main/cpp/llama-android.cpp b/examples/llama.android/llama/src/main/cpp/llama-android.cpp index 92a6b16b1..2aafe2316 100644 --- a/examples/llama.android/llama/src/main/cpp/llama-android.cpp +++ b/examples/llama.android/llama/src/main/cpp/llama-android.cpp @@ -409,7 +409,7 @@ Java_android_llama_cpp_LLamaAndroid_completion_1loop( const auto n_cur = env->CallIntMethod(intvar_ncur, la_int_var_value); if (llama_token_is_eog(model, new_token_id) || n_cur == n_len) { - return env->NewStringUTF(""); + return nullptr; } auto new_token_chars = llama_token_to_piece(context, new_token_id); diff --git a/examples/llama.swiftui/llama.cpp.swift/LibLlama.swift b/examples/llama.swiftui/llama.cpp.swift/LibLlama.swift index 737f882fb..58c32ca53 100644 --- a/examples/llama.swiftui/llama.cpp.swift/LibLlama.swift +++ b/examples/llama.swiftui/llama.cpp.swift/LibLlama.swift @@ -26,11 +26,12 @@ actor LlamaContext { private var context: OpaquePointer private var batch: llama_batch private var tokens_list: [llama_token] + var is_done: Bool = false /// This variable is used to store temporarily invalid cchars private var temporary_invalid_cchars: [CChar] - var n_len: Int32 = 64 + var n_len: Int32 = 1024 var n_cur: Int32 = 0 var n_decode: Int32 = 0 @@ -160,6 +161,7 @@ actor LlamaContext { if llama_token_is_eog(model, new_token_id) || n_cur == n_len { print("\n") + is_done = true let new_token_str = String(cString: temporary_invalid_cchars + [0]) temporary_invalid_cchars.removeAll() return new_token_str @@ -322,7 +324,7 @@ actor LlamaContext { defer { result.deallocate() } - let nTokens = llama_token_to_piece(model, token, result, 8, false) + let nTokens = llama_token_to_piece(model, token, result, 8, 0, false) if nTokens < 0 { let newResult = UnsafeMutablePointer.allocate(capacity: Int(-nTokens)) @@ -330,7 +332,7 @@ actor LlamaContext { defer { newResult.deallocate() } - let nNewTokens = llama_token_to_piece(model, token, newResult, -nTokens, false) + let nNewTokens = llama_token_to_piece(model, token, newResult, -nTokens, 0, false) let bufferPointer = UnsafeBufferPointer(start: newResult, count: Int(nNewTokens)) return Array(bufferPointer) } else { diff --git a/examples/llama.swiftui/llama.swiftui/Models/LlamaState.swift b/examples/llama.swiftui/llama.swiftui/Models/LlamaState.swift index 2c1e3f61b..b8f6a31d5 100644 --- a/examples/llama.swiftui/llama.swiftui/Models/LlamaState.swift +++ b/examples/llama.swiftui/llama.swiftui/Models/LlamaState.swift @@ -132,7 +132,7 @@ class LlamaState: ObservableObject { messageLog += "\(text)" Task.detached { - while await llamaContext.n_cur < llamaContext.n_len { + while await !llamaContext.is_done { let result = await llamaContext.completion_loop() await MainActor.run { self.messageLog += "\(result)" diff --git a/examples/llava/MobileVLM-README.md b/examples/llava/MobileVLM-README.md index f6c619c87..06a65fba4 100644 --- a/examples/llava/MobileVLM-README.md +++ b/examples/llava/MobileVLM-README.md @@ -30,16 +30,16 @@ git clone https://huggingface.co/mtgv/MobileVLM-1.7B git clone https://huggingface.co/openai/clip-vit-large-patch14-336 ``` -2. Use `llava-surgery.py` to split the LLaVA model to LLaMA and multimodel projector constituents: +2. Use `llava_surgery.py` to split the LLaVA model to LLaMA and multimodel projector constituents: ```sh -python ./examples/llava/llava-surgery.py -m path/to/MobileVLM-1.7B +python ./examples/llava/llava_surgery.py -m path/to/MobileVLM-1.7B ``` -3. Use `convert-image-encoder-to-gguf.py` with `--projector-type ldp` (for **V2** please use `--projector-type ldpv2`) to convert the LLaVA image encoder to GGUF: +3. Use `convert_image_encoder_to_gguf.py` with `--projector-type ldp` (for **V2** please use `--projector-type ldpv2`) to convert the LLaVA image encoder to GGUF: ```sh -python ./examples/llava/convert-image-encoder-to-gguf \ +python ./examples/llava/convert_image_encoder_to_gguf \ -m path/to/clip-vit-large-patch14-336 \ --llava-projector path/to/MobileVLM-1.7B/llava.projector \ --output-dir path/to/MobileVLM-1.7B \ @@ -47,17 +47,17 @@ python ./examples/llava/convert-image-encoder-to-gguf \ ``` ```sh -python ./examples/llava/convert-image-encoder-to-gguf \ +python ./examples/llava/convert_image_encoder_to_gguf \ -m path/to/clip-vit-large-patch14-336 \ --llava-projector path/to/MobileVLM-1.7B_V2/llava.projector \ --output-dir path/to/MobileVLM-1.7B_V2 \ --projector-type ldpv2 ``` -4. Use `examples/convert-legacy-llama.py` to convert the LLaMA part of LLaVA to GGUF: +4. Use `examples/convert_legacy_llama.py` to convert the LLaMA part of LLaVA to GGUF: ```sh -python ./examples/convert-legacy-llama.py path/to/MobileVLM-1.7B +python ./examples/convert_legacy_llama.py path/to/MobileVLM-1.7B ``` 5. Use `quantize` to convert LLaMA part's DataType from `fp16` to `q4_k` diff --git a/examples/llava/README.md b/examples/llava/README.md index f4554de67..012451361 100644 --- a/examples/llava/README.md +++ b/examples/llava/README.md @@ -38,22 +38,22 @@ git clone https://huggingface.co/openai/clip-vit-large-patch14-336 pip install -r examples/llava/requirements.txt ``` -3. Use `llava-surgery.py` to split the LLaVA model to LLaMA and multimodel projector constituents: +3. Use `llava_surgery.py` to split the LLaVA model to LLaMA and multimodel projector constituents: ```sh -python ./examples/llava/llava-surgery.py -m ../llava-v1.5-7b +python ./examples/llava/llava_surgery.py -m ../llava-v1.5-7b ``` -4. Use `convert-image-encoder-to-gguf.py` to convert the LLaVA image encoder to GGUF: +4. Use `convert_image_encoder_to_gguf.py` to convert the LLaVA image encoder to GGUF: ```sh -python ./examples/llava/convert-image-encoder-to-gguf.py -m ../clip-vit-large-patch14-336 --llava-projector ../llava-v1.5-7b/llava.projector --output-dir ../llava-v1.5-7b +python ./examples/llava/convert_image_encoder_to_gguf.py -m ../clip-vit-large-patch14-336 --llava-projector ../llava-v1.5-7b/llava.projector --output-dir ../llava-v1.5-7b ``` -5. Use `examples/convert-legacy-llama.py` to convert the LLaMA part of LLaVA to GGUF: +5. Use `examples/convert_legacy_llama.py` to convert the LLaMA part of LLaVA to GGUF: ```sh -python ./examples/convert-legacy-llama.py ../llava-v1.5-7b --skip-unknown +python ./examples/convert_legacy_llama.py ../llava-v1.5-7b --skip-unknown ``` Now both the LLaMA part and the image encoder are in the `llava-v1.5-7b` directory. @@ -70,9 +70,9 @@ git clone https://huggingface.co/liuhaotian/llava-v1.6-vicuna-7b pip install -r examples/llava/requirements.txt ``` -3) Use `llava-surgery-v2.py` which also supports llava-1.5 variants pytorch as well as safetensor models: +3) Use `llava_surgery_v2.py` which also supports llava-1.5 variants pytorch as well as safetensor models: ```console -python examples/llava/llava-surgery-v2.py -C -m ../llava-v1.6-vicuna-7b/ +python examples/llava/llava_surgery_v2.py -C -m ../llava-v1.6-vicuna-7b/ ``` - you will find a llava.projector and a llava.clip file in your model directory @@ -86,13 +86,13 @@ curl -s -q https://huggingface.co/cmp-nct/llava-1.6-gguf/raw/main/config_vit.jso 5) Create the visual gguf model: ```console -python ./examples/llava/convert-image-encoder-to-gguf.py -m vit --llava-projector vit/llava.projector --output-dir vit --clip-model-is-vision +python ./examples/llava/convert_image_encoder_to_gguf.py -m vit --llava-projector vit/llava.projector --output-dir vit --clip-model-is-vision ``` - This is similar to llava-1.5, the difference is that we tell the encoder that we are working with the pure vision model part of CLIP 6) Then convert the model to gguf format: ```console -python ./examples/convert-legacy-llama.py ../llava-v1.6-vicuna-7b/ --skip-unknown +python ./examples/convert_legacy_llama.py ../llava-v1.6-vicuna-7b/ --skip-unknown ``` 7) And finally we can run the llava cli using the 1.6 model version: diff --git a/examples/llava/clip.cpp b/examples/llava/clip.cpp index d6882eec3..7cda5f10c 100644 --- a/examples/llava/clip.cpp +++ b/examples/llava/clip.cpp @@ -16,6 +16,10 @@ #include "ggml-metal.h" #endif +#ifdef GGML_USE_CANN +#include "ggml-cann.h" +#endif + #define STB_IMAGE_IMPLEMENTATION #include "stb_image.h" @@ -865,7 +869,7 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32 embeddings = peg_0; } else { - GGML_ASSERT(false); + GGML_ABORT("fatal error"); } } @@ -1001,6 +1005,11 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) { LOG_TEE("%s: CLIP using Metal backend\n", __func__); #endif +#ifdef GGML_USE_CANN + new_clip->backend = ggml_backend_cann_init(0); + LOG_TEE("%s: CLIP using CANN backend\n", __func__); +#endif + if (!new_clip->backend) { new_clip->backend = ggml_backend_cpu_init(); diff --git a/examples/llava/convert-image-encoder-to-gguf.py b/examples/llava/convert_image_encoder_to_gguf.py similarity index 97% rename from examples/llava/convert-image-encoder-to-gguf.py rename to examples/llava/convert_image_encoder_to_gguf.py index b00bf7c6d..36f6b92fb 100644 --- a/examples/llava/convert-image-encoder-to-gguf.py +++ b/examples/llava/convert_image_encoder_to_gguf.py @@ -185,6 +185,8 @@ else: fout.add_description("two-tower CLIP model") if has_text_encoder: + assert t_hparams is not None + assert tokens is not None # text_model hparams fout.add_uint32(k(KEY_CONTEXT_LENGTH, TEXT), t_hparams["max_position_embeddings"]) fout.add_uint32(k(KEY_EMBEDDING_LENGTH, TEXT), t_hparams["hidden_size"]) @@ -259,8 +261,8 @@ if has_vision_encoder: if processor is not None: - image_mean = processor.image_processor.image_mean if args.image_mean is None or args.image_mean == default_image_mean else args.image_mean - image_std = processor.image_processor.image_std if args.image_std is None or args.image_std == default_image_std else args.image_std + image_mean = processor.image_processor.image_mean if args.image_mean is None or args.image_mean == default_image_mean else args.image_mean # pyright: ignore[reportAttributeAccessIssue] + image_std = processor.image_processor.image_std if args.image_std is None or args.image_std == default_image_std else args.image_std # pyright: ignore[reportAttributeAccessIssue] else: image_mean = args.image_mean if args.image_mean is not None else default_image_mean image_std = args.image_std if args.image_std is not None else default_image_std @@ -272,7 +274,7 @@ fout.add_bool("clip.use_gelu", use_gelu) if has_llava_projector: - model.vision_model.encoder.layers.pop(-1) + model.vision_model.encoder.layers.pop(-1) # pyright: ignore[reportAttributeAccessIssue] projector = torch.load(args.llava_projector) for name, data in projector.items(): name = get_tensor_name(name) @@ -286,7 +288,7 @@ if has_llava_projector: print("Projector tensors added\n") -state_dict = model.state_dict() +state_dict = model.state_dict() # pyright: ignore[reportAttributeAccessIssue] for name, data in state_dict.items(): if should_skip_tensor(name, has_text_encoder, has_vision_encoder, has_llava_projector): # we don't need this diff --git a/examples/llava/llava-surgery.py b/examples/llava/llava_surgery.py similarity index 100% rename from examples/llava/llava-surgery.py rename to examples/llava/llava_surgery.py diff --git a/examples/llava/llava-surgery-v2.py b/examples/llava/llava_surgery_v2.py similarity index 94% rename from examples/llava/llava-surgery-v2.py rename to examples/llava/llava_surgery_v2.py index eb56d6988..2d5b32fe6 100644 --- a/examples/llava/llava-surgery-v2.py +++ b/examples/llava/llava_surgery_v2.py @@ -2,7 +2,9 @@ import argparse import glob import os import torch -from safetensors.torch import load as safe_load, save as safe_save, safe_open, save_file +from safetensors import safe_open +from safetensors.torch import save_file +from typing import Any, ContextManager, cast # Function to determine if file is a SafeTensor file def is_safetensor_file(file_path): @@ -13,7 +15,7 @@ def is_safetensor_file(file_path): def load_model(file_path): if is_safetensor_file(file_path): tensors = {} - with safe_open(file_path, framework="pt", device="cpu") as f: + with cast(ContextManager[Any], safe_open(file_path, framework="pt", device="cpu")) as f: for key in f.keys(): tensors[key] = f.get_tensor(key).clone() # output shape @@ -134,7 +136,7 @@ if len(mm_tensors) == 0: if last_checkpoint is not None: for k, v in last_checkpoint.items(): print(k) - print(f"Found {len(mm_tensors)} tensors to extract out of {len(last_checkpoint)} tensors.") + print(f"Found {len(mm_tensors)} tensors to extract out of {len(last_checkpoint) if last_checkpoint is not None else 0} tensors.") print("No tensors found. Is this a LLaVA model?") exit() @@ -143,8 +145,10 @@ print(f"Found additional {len(first_mm_tensors)} tensors to extract.") # projector = {name: checkpoint.[name].float() for name in mm_tensors} projector = {} for name in mm_tensors: + assert last_checkpoint is not None projector[name] = last_checkpoint[name].float() for name in first_mm_tensors: + assert first_checkpoint is not None projector[name] = first_checkpoint[name].float() if len(projector) > 0: diff --git a/examples/llava/requirements.txt b/examples/llava/requirements.txt index 17cb4d5e5..fbfd0cdd7 100644 --- a/examples/llava/requirements.txt +++ b/examples/llava/requirements.txt @@ -1,3 +1,4 @@ --r ../../requirements/requirements-convert-legacy-llama.txt +-r ../../requirements/requirements-convert_legacy_llama.txt +--extra-index-url https://download.pytorch.org/whl/cpu pillow~=10.2.0 -torch~=2.1.1 +torch~=2.2.1 diff --git a/examples/lookup/README.md b/examples/lookup/README.md index 5bfb0de93..71c345c03 100644 --- a/examples/lookup/README.md +++ b/examples/lookup/README.md @@ -10,4 +10,3 @@ More info: https://github.com/ggerganov/llama.cpp/pull/4484 https://github.com/ggerganov/llama.cpp/issues/4226 - diff --git a/examples/lookup/lookup-stats.cpp b/examples/lookup/lookup-stats.cpp index 0b171c872..2fe67100e 100644 --- a/examples/lookup/lookup-stats.cpp +++ b/examples/lookup/lookup-stats.cpp @@ -31,7 +31,6 @@ int main(int argc, char ** argv){ // load the model std::tie(model, ctx) = llama_init_from_gpt_params(params); - GGML_ASSERT(llama_n_vocab(model) < (1 << 16)); // tokenize the prompt std::vector inp; @@ -65,7 +64,7 @@ int main(int argc, char ** argv){ } const int n_input = inp.size(); - const int n_ctx = params.n_ctx; + const int n_ctx = llama_n_ctx(ctx); int n_drafted = 0; int n_accept = 0; diff --git a/examples/lookup/lookup.cpp b/examples/lookup/lookup.cpp index 80ecd925d..bb571bac4 100644 --- a/examples/lookup/lookup.cpp +++ b/examples/lookup/lookup.cpp @@ -39,7 +39,6 @@ int main(int argc, char ** argv){ // load the model std::tie(model, ctx) = llama_init_from_gpt_params(params); - GGML_ASSERT(llama_n_vocab(model) < (1 << 16)); // tokenize the prompt std::vector inp; diff --git a/examples/main-cmake-pkg/.gitignore b/examples/main-cmake-pkg/.gitignore index e32c11c7f..67c01d64c 100644 --- a/examples/main-cmake-pkg/.gitignore +++ b/examples/main-cmake-pkg/.gitignore @@ -48,4 +48,3 @@ build*/ out/ tmp/ - diff --git a/examples/main-cmake-pkg/CMakeLists.txt b/examples/main-cmake-pkg/CMakeLists.txt index a97ded365..3b38db292 100644 --- a/examples/main-cmake-pkg/CMakeLists.txt +++ b/examples/main-cmake-pkg/CMakeLists.txt @@ -30,4 +30,3 @@ target_include_directories(${TARGET} PRIVATE ${_common_path}) install(TARGETS ${TARGET} RUNTIME) target_link_libraries(${TARGET} PRIVATE common llama ${CMAKE_THREAD_LIBS_INIT}) target_compile_features(${TARGET} PRIVATE cxx_std_11) - diff --git a/examples/main/README.md b/examples/main/README.md index 61e4a42f7..9396a34fa 100644 --- a/examples/main/README.md +++ b/examples/main/README.md @@ -1,6 +1,6 @@ # llama.cpp/examples/main -This example program allows you to use various LLaMA language models in an easy and efficient way. It is specifically designed to work with the [llama.cpp](https://github.com/ggerganov/llama.cpp) project, which provides a plain C/C++ implementation with optional 4-bit quantization support for faster, lower memory inference, and is optimized for desktop CPUs. This program can be used to perform various inference tasks with LLaMA models, including generating text based on user-provided prompts and chat-like interactions with reverse prompts. +This example program allows you to use various LLaMA language models easily and efficiently. It is specifically designed to work with the [llama.cpp](https://github.com/ggerganov/llama.cpp) project, which provides a plain C/C++ implementation with optional 4-bit quantization support for faster, lower memory inference, and is optimized for desktop CPUs. This program can be used to perform various inference tasks with LLaMA models, including generating text based on user-provided prompts and chat-like interactions with reverse prompts. ## Table of Contents @@ -17,60 +17,59 @@ This example program allows you to use various LLaMA language models in an easy To get started right away, run the following command, making sure to use the correct path for the model you have: -#### Unix-based systems (Linux, macOS, etc.): +First, we will need to download a model. In these examples, we will use the Gemma model from the ggml-org repo on Hugging Face. +[https://huggingface.co/ggml-org/gemma-1.1-7b-it-Q4_K_M-GGUF/resolve/main/gemma-1.1-7b-it.Q4_K_M.gguf?download=true](https://huggingface.co/ggml-org/gemma-1.1-7b-it-Q4_K_M-GGUF/resolve/main/gemma-1.1-7b-it.Q4_K_M.gguf?download=true) + +Once downloaded, place your model in the models folder in llama.cpp. + +### Unix-based systems (Linux, macOS, etc.): + +##### Input prompt (One-and-done) ```bash -./llama-cli -m models/7B/ggml-model.bin --prompt "Once upon a time" +./llama-cli -m models/gemma-1.1-7b-it.Q4_K_M.gguf --prompt "Once upon a time" ``` - -#### Windows: - -```powershell -llama-cli.exe -m models\7B\ggml-model.bin --prompt "Once upon a time" -``` - -For an interactive experience, try this command: - -#### Unix-based systems (Linux, macOS, etc.): +##### Conversation mode (Allow for continuous interaction with the model) ```bash -./llama-cli -m models/7B/ggml-model.bin -n -1 --color -r "User:" --in-prefix " " -i -p \ -'User: Hi -AI: Hello. I am an AI chatbot. Would you like to talk? -User: Sure! -AI: What would you like to talk about? -User:' +./llama-cli -m models/gemma-1.1-7b-it.Q4_K_M.gguf -cnv --chat-template gemma ``` -#### Windows: - -```powershell -llama-cli.exe -m models\7B\ggml-model.bin -n -1 --color -r "User:" --in-prefix " " -i -e -p "User: Hi\nAI: Hello. I am an AI chatbot. Would you like to talk?\nUser: Sure!\nAI: What would you like to talk about?\nUser:" -``` - -The following command generates "infinite" text from a starting prompt (you can use `Ctrl-C` to stop it): - -#### Unix-based systems (Linux, macOS, etc.): - +##### Infinite text from a starting prompt (you can use `Ctrl-C` to stop it): ```bash -./llama-cli -m models/7B/ggml-model.bin --ignore-eos -n -1 +./llama-cli -m models\gemma-1.1-7b-it.Q4_K_M.gguf --ignore-eos -n -1 ``` -#### Windows: +### Windows: + +##### Input prompt (One-and-done) +```powershell +./llama-cli.exe -m models\gemma-1.1-7b-it.Q4_K_M.gguf --prompt "Once upon a time" +``` +##### Conversation mode (Allow for continuous interaction with the model) ```powershell -llama-cli.exe -m models\7B\ggml-model.bin --ignore-eos -n -1 +./llama-cli.exe -m models\gemma-1.1-7b-it.Q4_K_M.gguf -cnv --chat-template gemma +``` + +#### Infinite text from a starting prompt (you can use `Ctrl-C` to stop it): + +```powershell +llama-cli.exe -m models\gemma-1.1-7b-it.Q4_K_M.gguf --ignore-eos -n -1 ``` ## Common Options In this section, we cover the most commonly used options for running the `llama-cli` program with the LLaMA models: -- `-m FNAME, --model FNAME`: Specify the path to the LLaMA model file (e.g., `models/7B/ggml-model.gguf`; inferred from `--model-url` if set). -- `-mu MODEL_URL --model-url MODEL_URL`: Specify a remote http url to download the file (e.g https://huggingface.co/ggml-org/models/resolve/main/phi-2/ggml-model-q4_0.gguf). +- `-m FNAME, --model FNAME`: Specify the path to the LLaMA model file (e.g., `models/gemma-1.1-7b-it.Q4_K_M.gguf`; inferred from `--model-url` if set). +- `-mu MODEL_URL --model-url MODEL_URL`: Specify a remote http url to download the file (e.g [https://huggingface.co/ggml-org/gemma-1.1-7b-it-Q4_K_M-GGUF/resolve/main/gemma-1.1-7b-it.Q4_K_M.gguf?download=true](https://huggingface.co/ggml-org/gemma-1.1-7b-it-Q4_K_M-GGUF/resolve/main/gemma-1.1-7b-it.Q4_K_M.gguf?download=true)). - `-i, --interactive`: Run the program in interactive mode, allowing you to provide input directly and receive real-time responses. - `-n N, --n-predict N`: Set the number of tokens to predict when generating text. Adjusting this value can influence the length of the generated text. - `-c N, --ctx-size N`: Set the size of the prompt context. The default is 512, but LLaMA models were built with a context of 2048, which will provide better results for longer input/inference. +- `-mli, --multiline-input`: Allows you to write or paste multiple lines without ending each in '\' +- `-t N, --threads N`: Set the number of threads to use during generation. For optimal performance, it is recommended to set this value to the number of physical CPU cores your system has. +- - `-ngl N, --n-gpu-layers N`: When compiled with GPU support, this option allows offloading some layers to the GPU for computation. Generally results in increased performance. ## Input Prompts @@ -90,6 +89,7 @@ In interactive mode, users can participate in text generation by injecting their - `-i, --interactive`: Run the program in interactive mode, allowing users to engage in real-time conversations or provide specific instructions to the model. - `--interactive-first`: Run the program in interactive mode and immediately wait for user input before starting the text generation. +- `-cnv, --conversation`: Run the program in conversation mode (does not print special tokens and suffix/prefix, use default chat template) (default: false) - `--color`: Enable colorized output to differentiate visually distinguishing between prompts, user input, and generated text. By understanding and utilizing these interaction options, you can create engaging and dynamic experiences with the LLaMA models, tailoring the text generation process to your specific needs. @@ -117,6 +117,13 @@ The `--in-suffix` flag is used to add a suffix after your input. This is useful ```sh ./llama-cli -r "User:" --in-prefix " " --in-suffix "Assistant:" ``` +When --in-prefix or --in-suffix options are enabled the chat template ( --chat-template ) is disabled + +### Chat templates + + `--chat-template JINJA_TEMPLATE`: This option sets a custom jinja chat template. It accepts a string, not a file name. Default: template taken from model's metadata. Llama.cpp only supports [some pre-defined templates](https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template). These include llama2, llama3, gemma, monarch, chatml, orion, vicuna, vicuna-orca, deepseek, command-r, zephyr. When --in-prefix or --in-suffix options are enabled the chat template ( --chat-template ) is disabled. + + Example usage: `--chat-template gemma` ## Context Management @@ -124,9 +131,7 @@ During text generation, LLaMA models have a limited context size, which means th ### Context Size -The `--ctx-size` option allows you to set the size of the prompt context used by the LLaMA models during text generation. A larger context size helps the model to better comprehend and generate responses for longer input or conversations. - -- `-c N, --ctx-size N`: Set the size of the prompt context (default: 512). The LLaMA models were built with a context of 2048, which will yield the best results on longer input/inference. However, increasing the context size beyond 2048 may lead to unpredictable results. +- `-c N, --ctx-size N`: Set the size of the prompt context (default: 0, 0 = loaded from model). The LLaMA models were built with a context of 2048-8192, which will yield the best results on longer input/inference. ### Extended Context Size @@ -148,15 +153,15 @@ The following options allow you to control the text generation process and fine- ### Number of Tokens to Predict -- `-n N, --n-predict N`: Set the number of tokens to predict when generating text (default: 128, -1 = infinity, -2 = until context filled) +- `-n N, --predict N`: Set the number of tokens to predict when generating text (default: -1, -1 = infinity, -2 = until context filled) -The `--n-predict` option controls the number of tokens the model generates in response to the input prompt. By adjusting this value, you can influence the length of the generated text. A higher value will result in longer text, while a lower value will produce shorter text. +The `--predict` option controls the number of tokens the model generates in response to the input prompt. By adjusting this value, you can influence the length of the generated text. A higher value will result in longer text, while a lower value will produce shorter text. -A value of -1 will enable infinite text generation, even though we have a finite context window. When the context window is full, some of the earlier tokens (half of the tokens after `--n-keep`) will be discarded. The context must then be re-evaluated before generation can resume. On large models and/or large context windows, this will result in significant pause in output. +A value of -1 will enable infinite text generation, even though we have a finite context window. When the context window is full, some of the earlier tokens (half of the tokens after `--keep`) will be discarded. The context must then be re-evaluated before generation can resume. On large models and/or large context windows, this will result in a significant pause in output. If the pause is undesirable, a value of -2 will stop generation immediately when the context is filled. -It is important to note that the generated text may be shorter than the specified number of tokens if an End-of-Sequence (EOS) token or a reverse prompt is encountered. In interactive mode text generation will pause and control will be returned to the user. In non-interactive mode, the program will end. In both cases, the text generation may stop before reaching the specified `n-predict` value. If you want the model to keep going without ever producing End-of-Sequence on its own, you can use the `--ignore-eos` parameter. +It is important to note that the generated text may be shorter than the specified number of tokens if an End-of-Sequence (EOS) token or a reverse prompt is encountered. In interactive mode, text generation will pause and control will be returned to the user. In non-interactive mode, the program will end. In both cases, the text generation may stop before reaching the specified `--predict` value. If you want the model to keep going without ever producing End-of-Sequence on its own, you can use the `--ignore-eos` parameter. ### Temperature @@ -164,15 +169,15 @@ It is important to note that the generated text may be shorter than the specifie Temperature is a hyperparameter that controls the randomness of the generated text. It affects the probability distribution of the model's output tokens. A higher temperature (e.g., 1.5) makes the output more random and creative, while a lower temperature (e.g., 0.5) makes the output more focused, deterministic, and conservative. The default value is 0.8, which provides a balance between randomness and determinism. At the extreme, a temperature of 0 will always pick the most likely next token, leading to identical outputs in each run. -Example usage: `--temp 0.5` +Example usage: `--temp 0` ### Repeat Penalty -- `--repeat-penalty N`: Control the repetition of token sequences in the generated text (default: 1.1). +- `--repeat-penalty N`: Control the repetition of token sequences in the generated text default: 1.0, 1.0 = disabled). - `--repeat-last-n N`: Last n tokens to consider for penalizing repetition (default: 64, 0 = disabled, -1 = ctx-size). - `--no-penalize-nl`: Disable penalization for newline tokens when applying the repeat penalty. -The `repeat-penalty` option helps prevent the model from generating repetitive or monotonous text. A higher value (e.g., 1.5) will penalize repetitions more strongly, while a lower value (e.g., 0.9) will be more lenient. The default value is 1.1. +The `repeat-penalty` option helps prevent the model from generating repetitive or monotonous text. A higher value (e.g., 1.5) will penalize repetitions more strongly, while a lower value (e.g., 0.9) will be more lenient. The default value is 1. The `repeat-last-n` option controls the number of tokens in the history to consider for penalizing repetition. A larger value will look further back in the generated text to prevent repetitions, while a smaller value will only consider recent tokens. A value of 0 disables the penalty, and a value of -1 sets the number of tokens considered equal to the context size (`ctx-size`). @@ -196,19 +201,19 @@ Top-p sampling, also known as nucleus sampling, is another text generation metho Example usage: `--top-p 0.95` -### Min P Sampling +### Min-P Sampling -- `--min-p N`: Sets a minimum base probability threshold for token selection (default: 0.05). +- `--min-p N`: Sets a minimum base probability threshold for token selection (default: 0.1). The Min-P sampling method was designed as an alternative to Top-P, and aims to ensure a balance of quality and variety. The parameter *p* represents the minimum probability for a token to be considered, relative to the probability of the most likely token. For example, with *p*=0.05 and the most likely token having a probability of 0.9, logits with a value less than 0.045 are filtered out. Example usage: `--min-p 0.05` -### Tail Free Sampling (TFS) +### Tail-Free Sampling (TFS) - `--tfs N`: Enable tail free sampling with parameter z (default: 1.0, 1.0 = disabled). -Tail free sampling (TFS) is a text generation technique that aims to reduce the impact of less likely tokens, which may be less relevant, less coherent, or nonsensical, on the output. Similar to Top-P it tries to determine the bulk of the most likely tokens dynamically. But TFS filters out logits based on the second derivative of their probabilities. Adding tokens is stopped after the sum of the second derivatives reaches the parameter z. In short: TFS looks how quickly the probabilities of the tokens decrease and cuts off the tail of unlikely tokens using the parameter z. Typical values for z are in the range of 0.9 to 0.95. A value of 1.0 would include all tokens, and thus disables the effect of TFS. +Tail-free sampling (TFS) is a text generation technique that aims to reduce the impact of less likely tokens, which may be less relevant, less coherent, or nonsensical, on the output. Similar to Top-P it tries to determine the bulk of the most likely tokens dynamically. But TFS filters out logits based on the second derivative of their probabilities. Adding tokens is stopped after the sum of the second derivatives reaches the parameter z. In short: TFS looks at how quickly the probabilities of the tokens decrease and cuts off the tail of unlikely tokens using the parameter z. Typical values for z are in the range of 0.9 to 0.95. A value of 1.0 would include all tokens and thus disables the effect of TFS. Example usage: `--tfs 0.95` @@ -307,10 +312,8 @@ These options provide extra functionality and customization when running the LLa - `-h, --help`: Display a help message showing all available options and their default values. This is particularly useful for checking the latest options and default values, as they can change frequently, and the information in this document may become outdated. - `--verbose-prompt`: Print the prompt before generating text. -- `-ngl N, --n-gpu-layers N`: When compiled with GPU support, this option allows offloading some layers to the GPU for computation. Generally results in increased performance. - `-mg i, --main-gpu i`: When using multiple GPUs this option controls which GPU is used for small tensors for which the overhead of splitting the computation across all GPUs is not worthwhile. The GPU in question will use slightly more VRAM to store a scratch buffer for temporary results. By default GPU 0 is used. - `-ts SPLIT, --tensor-split SPLIT`: When using multiple GPUs this option controls how large tensors should be split across all GPUs. `SPLIT` is a comma-separated list of non-negative values that assigns the proportion of data that each GPU should get in order. For example, "3,2" will assign 60% of the data to GPU 0 and 40% to GPU 1. By default the data is split in proportion to VRAM but this may not be optimal for performance. - `--lora FNAME`: Apply a LoRA (Low-Rank Adaptation) adapter to the model (implies --no-mmap). This allows you to adapt the pretrained model to specific tasks or domains. - `--lora-base FNAME`: Optional model to use as a base for the layers modified by the LoRA adapter. This flag is used in conjunction with the `--lora` flag, and specifies the base model for the adaptation. - - `-hfr URL --hf-repo URL`: The url to the Hugging Face model repository. Used in conjunction with `--hf-file` or `-hff`. The model is downloaded and stored in the file provided by `-m` or `--model`. If `-m` is not provided, the model is auto-stored in the path specified by the `LLAMA_CACHE` environment variable or in an OS-specific local cache. diff --git a/examples/main/main.cpp b/examples/main/main.cpp index cfaf6a6e8..61e960ea2 100644 --- a/examples/main/main.cpp +++ b/examples/main/main.cpp @@ -37,7 +37,8 @@ static gpt_params * g_params; static std::vector * g_input_tokens; static std::ostringstream * g_output_ss; static std::vector * g_output_tokens; -static bool is_interacting = false; +static bool is_interacting = false; +static bool need_insert_eot = false; static bool file_exists(const std::string & path) { std::ifstream f(path.c_str()); @@ -99,7 +100,8 @@ static void write_logfile( static void sigint_handler(int signo) { if (signo == SIGINT) { if (!is_interacting && g_params->interactive) { - is_interacting = true; + is_interacting = true; + need_insert_eot = true; } else { console::cleanup(); printf("\n"); @@ -122,6 +124,7 @@ static std::string chat_add_and_format(struct llama_model * model, std::vectorchat_template, chat_msgs, new_msg, role == "user"); chat_msgs.push_back({role, content}); + LOG("formatted: %s\n", formatted.c_str()); return formatted; } @@ -224,7 +227,14 @@ int main(int argc, char ** argv) { __func__, n_ctx_train, n_ctx); } - LOG_TEE("%s: chat template example: %s\n", __func__, llama_chat_format_example(model, params.chat_template).c_str()); + // print chat template example in conversation mode + if (params.conversation) { + if (params.enable_chat_template) { + LOG_TEE("%s: chat template example: %s\n", __func__, llama_chat_format_example(model, params.chat_template).c_str()); + } else { + LOG_TEE("%s: in-suffix/prefix is specified, chat template will be disabled\n", __func__); + } + } // print system information { @@ -255,13 +265,15 @@ int main(int argc, char ** argv) { } const bool add_bos = llama_should_add_bos_token(model); - GGML_ASSERT(llama_add_eos_token(model) != 1); + if (!llama_model_has_encoder(model)) { + GGML_ASSERT(llama_add_eos_token(model) != 1); + } LOG("add_bos: %d\n", add_bos); std::vector embd_inp; { - auto prompt = params.conversation + auto prompt = (params.conversation && params.enable_chat_template && !params.prompt.empty()) ? chat_add_and_format(model, chat_msgs, "system", params.prompt) // format the system prompt in conversation mode : params.prompt; if (params.interactive_first || !params.prompt.empty() || session_tokens.empty()) { @@ -278,8 +290,13 @@ int main(int argc, char ** argv) { // Should not run without any tokens if (embd_inp.empty()) { - embd_inp.push_back(llama_token_bos(model)); - LOG("embd_inp was considered empty and bos was added: %s\n", LOG_TOKENS_TOSTR_PRETTY(ctx, embd_inp).c_str()); + if (add_bos) { + embd_inp.push_back(llama_token_bos(model)); + LOG("embd_inp was considered empty and bos was added: %s\n", LOG_TOKENS_TOSTR_PRETTY(ctx, embd_inp).c_str()); + } else { + LOG_TEE("error: input is empty\n"); + return -1; + } } // Tokenize negative prompt @@ -517,6 +534,24 @@ int main(int argc, char ** argv) { exit(1); } + if (llama_model_has_encoder(model)) { + int enc_input_size = embd_inp.size(); + llama_token * enc_input_buf = embd_inp.data(); + + if (llama_encode(ctx, llama_batch_get_one(enc_input_buf, enc_input_size, 0, 0))) { + LOG_TEE("%s : failed to eval\n", __func__); + return 1; + } + + llama_token decoder_start_token_id = llama_model_decoder_start_token(model); + if (decoder_start_token_id == -1) { + decoder_start_token_id = llama_token_bos(model); + } + + embd_inp.clear(); + embd_inp.push_back(decoder_start_token_id); + } + while ((n_remain != 0 && !is_antiprompt) || params.interactive) { // predict if (!embd.empty()) { @@ -810,7 +845,9 @@ int main(int argc, char ** argv) { is_antiprompt = true; } - chat_add_and_format(model, chat_msgs, "system", assistant_ss.str()); + if (params.enable_chat_template) { + chat_add_and_format(model, chat_msgs, "assistant", assistant_ss.str()); + } is_interacting = true; printf("\n"); } @@ -872,16 +909,24 @@ int main(int argc, char ** argv) { string_process_escapes(buffer); } - std::string user_inp = params.conversation + bool format_chat = params.conversation && params.enable_chat_template; + std::string user_inp = format_chat ? chat_add_and_format(model, chat_msgs, "user", std::move(buffer)) : std::move(buffer); // TODO: one inconvenient of current chat template implementation is that we can't distinguish between user input and special tokens (prefix/postfix) const auto line_pfx = ::llama_tokenize(ctx, params.input_prefix, false, true); - const auto line_inp = ::llama_tokenize(ctx, user_inp, false, params.conversation); + const auto line_inp = ::llama_tokenize(ctx, user_inp, false, format_chat); const auto line_sfx = ::llama_tokenize(ctx, params.input_suffix, false, true); LOG("input tokens: %s\n", LOG_TOKENS_TOSTR_PRETTY(ctx, line_inp).c_str()); + // if user stop generation mid-way, we must add EOT to finish model's last response + if (need_insert_eot && format_chat) { + llama_token eot = llama_token_eot(model); + embd_inp.push_back(eot == -1 ? llama_token_eos(model) : eot); + need_insert_eot = false; + } + embd_inp.insert(embd_inp.end(), line_pfx.begin(), line_pfx.end()); embd_inp.insert(embd_inp.end(), line_inp.begin(), line_inp.end()); embd_inp.insert(embd_inp.end(), line_sfx.begin(), line_sfx.end()); diff --git a/examples/passkey/README.md b/examples/passkey/README.md index a48a6283a..2b8e910f9 100644 --- a/examples/passkey/README.md +++ b/examples/passkey/README.md @@ -1,5 +1,8 @@ # llama.cpp/example/passkey +A passkey retrieval task is an evaluation method used to measure a language +models ability to recall information from long contexts. + See the following PRs for more info: - https://github.com/ggerganov/llama.cpp/pull/3856 diff --git a/examples/perplexity/perplexity.cpp b/examples/perplexity/perplexity.cpp index efde8dfdf..dbe445391 100644 --- a/examples/perplexity/perplexity.cpp +++ b/examples/perplexity/perplexity.cpp @@ -1991,6 +1991,12 @@ int main(int argc, char ** argv) { params.n_batch = std::min(params.n_batch, n_kv); } else { params.n_batch = std::min(params.n_batch, params.n_ctx); + if (params.kl_divergence) { + params.n_parallel = 1; + } else { + // ensure there's at least enough seq_ids for HellaSwag + params.n_parallel = std::max(4, params.n_parallel); + } } if (params.ppl_stride > 0) { @@ -2015,9 +2021,6 @@ int main(int argc, char ** argv) { llama_model * model; llama_context * ctx; - // ensure there's at least enough seq_ids for HellaSwag - params.n_parallel = std::max(4, params.n_parallel); - // load the model and apply lora adapter, if any std::tie(model, ctx) = llama_init_from_gpt_params(params); if (model == NULL) { diff --git a/examples/pydantic-models-to-grammar-examples.py b/examples/pydantic-models-to-grammar-examples.py deleted file mode 100644 index 160966649..000000000 --- a/examples/pydantic-models-to-grammar-examples.py +++ /dev/null @@ -1,224 +0,0 @@ -# Function calling example using pydantic models. -import datetime -import importlib -import json -from enum import Enum -from typing import Optional, Union - -import requests -from pydantic import BaseModel, Field -from pydantic_models_to_grammar import (add_run_method_to_dynamic_model, convert_dictionary_to_pydantic_model, - create_dynamic_model_from_function, generate_gbnf_grammar_and_documentation) - - -# Function to get completion on the llama.cpp server with grammar. -def create_completion(prompt, grammar): - headers = {"Content-Type": "application/json"} - data = {"prompt": prompt, "grammar": grammar} - - response = requests.post("http://127.0.0.1:8080/completion", headers=headers, json=data) - data = response.json() - - print(data["content"]) - return data["content"] - - -# A function for the agent to send a message to the user. -class SendMessageToUser(BaseModel): - """ - Send a message to the User. - """ - chain_of_thought: str = Field(..., description="Your chain of thought while sending the message.") - message: str = Field(..., description="Message you want to send to the user.") - - def run(self): - print(self.message) - - -# Enum for the calculator tool. -class MathOperation(Enum): - ADD = "add" - SUBTRACT = "subtract" - MULTIPLY = "multiply" - DIVIDE = "divide" - - -# Simple pydantic calculator tool for the agent that can add, subtract, multiply, and divide. Docstring and description of fields will be used in system prompt. -class Calculator(BaseModel): - """ - Perform a math operation on two numbers. - """ - number_one: Union[int, float] = Field(..., description="First number.") - operation: MathOperation = Field(..., description="Math operation to perform.") - number_two: Union[int, float] = Field(..., description="Second number.") - - def run(self): - if self.operation == MathOperation.ADD: - return self.number_one + self.number_two - elif self.operation == MathOperation.SUBTRACT: - return self.number_one - self.number_two - elif self.operation == MathOperation.MULTIPLY: - return self.number_one * self.number_two - elif self.operation == MathOperation.DIVIDE: - return self.number_one / self.number_two - else: - raise ValueError("Unknown operation.") - - -# Here the grammar gets generated by passing the available function models to generate_gbnf_grammar_and_documentation function. This also generates a documentation usable by the LLM. -# pydantic_model_list is the list of pydanitc models -# outer_object_name is an optional name for an outer object around the actual model object. Like a "function" object with "function_parameters" which contains the actual model object. If None, no outer object will be generated -# outer_object_content is the name of outer object content. -# model_prefix is the optional prefix for models in the documentation. (Default="Output Model") -# fields_prefix is the prefix for the model fields in the documentation. (Default="Output Fields") -gbnf_grammar, documentation = generate_gbnf_grammar_and_documentation( - pydantic_model_list=[SendMessageToUser, Calculator], outer_object_name="function", - outer_object_content="function_parameters", model_prefix="Function", fields_prefix="Parameters") - -print(gbnf_grammar) -print(documentation) - -system_message = "You are an advanced AI, tasked to assist the user by calling functions in JSON format. The following are the available functions and their parameters and types:\n\n" + documentation - -user_message = "What is 42 * 42?" -prompt = f"<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{user_message}<|im_end|>\n<|im_start|>assistant" - -text = create_completion(prompt=prompt, grammar=gbnf_grammar) -# This should output something like this: -# { -# "function": "calculator", -# "function_parameters": { -# "number_one": 42, -# "operation": "multiply", -# "number_two": 42 -# } -# } -function_dictionary = json.loads(text) -if function_dictionary["function"] == "calculator": - function_parameters = {**function_dictionary["function_parameters"]} - - print(Calculator(**function_parameters).run()) - # This should output: 1764 - - -# A example structured output based on pydantic models. The LLM will create an entry for a Book database out of an unstructured text. -class Category(Enum): - """ - The category of the book. - """ - Fiction = "Fiction" - NonFiction = "Non-Fiction" - - -class Book(BaseModel): - """ - Represents an entry about a book. - """ - title: str = Field(..., description="Title of the book.") - author: str = Field(..., description="Author of the book.") - published_year: Optional[int] = Field(..., description="Publishing year of the book.") - keywords: list[str] = Field(..., description="A list of keywords.") - category: Category = Field(..., description="Category of the book.") - summary: str = Field(..., description="Summary of the book.") - - -# We need no additional parameters other than our list of pydantic models. -gbnf_grammar, documentation = generate_gbnf_grammar_and_documentation([Book]) - -system_message = "You are an advanced AI, tasked to create a dataset entry in JSON for a Book. The following is the expected output model:\n\n" + documentation - -text = """The Feynman Lectures on Physics is a physics textbook based on some lectures by Richard Feynman, a Nobel laureate who has sometimes been called "The Great Explainer". The lectures were presented before undergraduate students at the California Institute of Technology (Caltech), during 1961–1963. The book's co-authors are Feynman, Robert B. Leighton, and Matthew Sands.""" -prompt = f"<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{text}<|im_end|>\n<|im_start|>assistant" - -text = create_completion(prompt=prompt, grammar=gbnf_grammar) - -json_data = json.loads(text) - -print(Book(**json_data)) -# An example for parallel function calling with a Python function, a pydantic function model and an OpenAI like function definition. - -def get_current_datetime(output_format: Optional[str] = None): - """ - Get the current date and time in the given format. - Args: - output_format: formatting string for the date and time, defaults to '%Y-%m-%d %H:%M:%S' - """ - if output_format is None: - output_format = '%Y-%m-%d %H:%M:%S' - return datetime.datetime.now().strftime(output_format) - - -# Example function to get the weather -def get_current_weather(location, unit): - """Get the current weather in a given location""" - if "London" in location: - return json.dumps({"location": "London", "temperature": "42", "unit": unit.value}) - elif "New York" in location: - return json.dumps({"location": "New York", "temperature": "24", "unit": unit.value}) - elif "North Pole" in location: - return json.dumps({"location": "North Pole", "temperature": "-42", "unit": unit.value}) - else: - return json.dumps({"location": location, "temperature": "unknown"}) - - -# Here is a function definition in OpenAI style -current_weather_tool = { - "type": "function", - "function": { - "name": "get_current_weather", - "description": "Get the current weather in a given location", - "parameters": { - "type": "object", - "properties": { - "location": { - "type": "string", - "description": "The city and state, e.g. San Francisco, CA", - }, - "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]}, - }, - "required": ["location"], - }, - }, -} - -# Convert OpenAI function definition into pydantic model -current_weather_tool_model = convert_dictionary_to_pydantic_model(current_weather_tool) -# Add the actual function to a pydantic model -current_weather_tool_model = add_run_method_to_dynamic_model(current_weather_tool_model, get_current_weather) - -# Convert normal Python function to a pydantic model -current_datetime_model = create_dynamic_model_from_function(get_current_datetime) - -tool_list = [SendMessageToUser, Calculator, current_datetime_model, current_weather_tool_model] - - -gbnf_grammar, documentation = generate_gbnf_grammar_and_documentation( - pydantic_model_list=tool_list, outer_object_name="function", - outer_object_content="params", model_prefix="Function", fields_prefix="Parameters", list_of_outputs=True) - -system_message = "You are an advanced AI assistant. You are interacting with the user and with your environment by calling functions. You call functions by writing JSON objects, which represent specific function calls.\nBelow is a list of your available function calls:\n\n" + documentation - - -text = """Get the date and time, get the current weather in celsius in London and solve the following calculation: 42 * 42""" -prompt = f"<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{text}<|im_end|>\n<|im_start|>assistant" - -text = create_completion(prompt=prompt, grammar=gbnf_grammar) - -json_data = json.loads(text) - -print(json_data) -# Should output something like this: -# [{'function': 'get_current_datetime', 'params': {'output_format': '%Y-%m-%d %H:%M:%S'}}, {'function': 'get_current_weather', 'params': {'location': 'London', 'unit': 'celsius'}}, {'function': 'Calculator', 'params': {'number_one': 42, 'operation': 'multiply', 'number_two': 42}}] - - -for call in json_data: - if call["function"] == "Calculator": - print(Calculator(**call["params"]).run()) - elif call["function"] == "get_current_datetime": - print(current_datetime_model(**call["params"]).run()) - elif call["function"] == "get_current_weather": - print(current_weather_tool_model(**call["params"]).run()) -# Should output something like this: -# 2024-01-14 13:36:06 -# {"location": "London", "temperature": "42", "unit": "celsius"} -# 1764 diff --git a/examples/pydantic_models_to_grammar.py b/examples/pydantic_models_to_grammar.py index f029c73a2..93e5dcb6c 100644 --- a/examples/pydantic_models_to_grammar.py +++ b/examples/pydantic_models_to_grammar.py @@ -9,7 +9,7 @@ from inspect import getdoc, isclass from typing import TYPE_CHECKING, Any, Callable, List, Optional, Union, get_args, get_origin, get_type_hints from docstring_parser import parse -from pydantic import BaseModel, Field, create_model +from pydantic import BaseModel, create_model if TYPE_CHECKING: from types import GenericAlias @@ -17,6 +17,9 @@ else: # python 3.8 compat from typing import _GenericAlias as GenericAlias +# TODO: fix this +# pyright: reportAttributeAccessIssue=information + class PydanticDataType(Enum): """ @@ -50,35 +53,38 @@ class PydanticDataType(Enum): def map_pydantic_type_to_gbnf(pydantic_type: type[Any]) -> str: - if isclass(pydantic_type) and issubclass(pydantic_type, str): + origin_type = get_origin(pydantic_type) + origin_type = pydantic_type if origin_type is None else origin_type + + if isclass(origin_type) and issubclass(origin_type, str): return PydanticDataType.STRING.value - elif isclass(pydantic_type) and issubclass(pydantic_type, bool): + elif isclass(origin_type) and issubclass(origin_type, bool): return PydanticDataType.BOOLEAN.value - elif isclass(pydantic_type) and issubclass(pydantic_type, int): + elif isclass(origin_type) and issubclass(origin_type, int): return PydanticDataType.INTEGER.value - elif isclass(pydantic_type) and issubclass(pydantic_type, float): + elif isclass(origin_type) and issubclass(origin_type, float): return PydanticDataType.FLOAT.value - elif isclass(pydantic_type) and issubclass(pydantic_type, Enum): + elif isclass(origin_type) and issubclass(origin_type, Enum): return PydanticDataType.ENUM.value - elif isclass(pydantic_type) and issubclass(pydantic_type, BaseModel): - return format_model_and_field_name(pydantic_type.__name__) - elif get_origin(pydantic_type) is list: + elif isclass(origin_type) and issubclass(origin_type, BaseModel): + return format_model_and_field_name(origin_type.__name__) + elif origin_type is list: element_type = get_args(pydantic_type)[0] return f"{map_pydantic_type_to_gbnf(element_type)}-list" - elif get_origin(pydantic_type) is set: + elif origin_type is set: element_type = get_args(pydantic_type)[0] return f"{map_pydantic_type_to_gbnf(element_type)}-set" - elif get_origin(pydantic_type) is Union: + elif origin_type is Union: union_types = get_args(pydantic_type) union_rules = [map_pydantic_type_to_gbnf(ut) for ut in union_types] return f"union-{'-or-'.join(union_rules)}" - elif get_origin(pydantic_type) is Optional: + elif origin_type is Optional: element_type = get_args(pydantic_type)[0] return f"optional-{map_pydantic_type_to_gbnf(element_type)}" - elif isclass(pydantic_type): - return f"{PydanticDataType.CUSTOM_CLASS.value}-{format_model_and_field_name(pydantic_type.__name__)}" - elif get_origin(pydantic_type) is dict: + elif isclass(origin_type): + return f"{PydanticDataType.CUSTOM_CLASS.value}-{format_model_and_field_name(origin_type.__name__)}" + elif origin_type is dict: key_type, value_type = get_args(pydantic_type) return f"custom-dict-key-type-{format_model_and_field_name(map_pydantic_type_to_gbnf(key_type))}-value-type-{format_model_and_field_name(map_pydantic_type_to_gbnf(value_type))}" else: @@ -115,7 +121,7 @@ def get_members_structure(cls, rule_name): # Modify this comprehension members = [ f' "\\"{name}\\"" ":" {map_pydantic_type_to_gbnf(param_type)}' - for name, param_type in cls.__annotations__.items() + for name, param_type in get_type_hints(cls).items() if name != "self" ] @@ -234,8 +240,9 @@ def generate_gbnf_float_rules(max_digit=None, min_digit=None, max_precision=None # Define the integer part rule integer_part_rule = ( - "integer-part" + (f"-max{max_digit}" if max_digit is not None else "") + ( - f"-min{min_digit}" if min_digit is not None else "") + "integer-part" + + (f"-max{max_digit}" if max_digit is not None else "") + + (f"-min{min_digit}" if min_digit is not None else "") ) # Define the fractional part rule based on precision constraints @@ -293,17 +300,20 @@ def generate_gbnf_rule_for_type( field_name = format_model_and_field_name(field_name) gbnf_type = map_pydantic_type_to_gbnf(field_type) - if isclass(field_type) and issubclass(field_type, BaseModel): + origin_type = get_origin(field_type) + origin_type = field_type if origin_type is None else origin_type + + if isclass(origin_type) and issubclass(origin_type, BaseModel): nested_model_name = format_model_and_field_name(field_type.__name__) nested_model_rules, _ = generate_gbnf_grammar(field_type, processed_models, created_rules) rules.extend(nested_model_rules) gbnf_type, rules = nested_model_name, rules - elif isclass(field_type) and issubclass(field_type, Enum): + elif isclass(origin_type) and issubclass(origin_type, Enum): enum_values = [f'"\\"{e.value}\\""' for e in field_type] # Adding escaped quotes enum_rule = f"{model_name}-{field_name} ::= {' | '.join(enum_values)}" rules.append(enum_rule) gbnf_type, rules = model_name + "-" + field_name, rules - elif get_origin(field_type) == list: # Array + elif origin_type is list: # Array element_type = get_args(field_type)[0] element_rule_name, additional_rules = generate_gbnf_rule_for_type( model_name, f"{field_name}-element", element_type, is_optional, processed_models, created_rules @@ -313,7 +323,7 @@ def generate_gbnf_rule_for_type( rules.append(array_rule) gbnf_type, rules = model_name + "-" + field_name, rules - elif get_origin(field_type) == set or field_type == set: # Array + elif origin_type is set: # Array element_type = get_args(field_type)[0] element_rule_name, additional_rules = generate_gbnf_rule_for_type( model_name, f"{field_name}-element", element_type, is_optional, processed_models, created_rules @@ -367,7 +377,7 @@ def generate_gbnf_rule_for_type( gbnf_type = f"{model_name}-{field_name}-optional" else: gbnf_type = f"{model_name}-{field_name}-union" - elif isclass(field_type) and issubclass(field_type, str): + elif isclass(origin_type) and issubclass(origin_type, str): if field_info and hasattr(field_info, "json_schema_extra") and field_info.json_schema_extra is not None: triple_quoted_string = field_info.json_schema_extra.get("triple_quoted_string", False) markdown_string = field_info.json_schema_extra.get("markdown_code_block", False) @@ -383,8 +393,8 @@ def generate_gbnf_rule_for_type( gbnf_type = PydanticDataType.STRING.value elif ( - isclass(field_type) - and issubclass(field_type, float) + isclass(origin_type) + and issubclass(origin_type, float) and field_info and hasattr(field_info, "json_schema_extra") and field_info.json_schema_extra is not None @@ -409,8 +419,8 @@ def generate_gbnf_rule_for_type( ) elif ( - isclass(field_type) - and issubclass(field_type, int) + isclass(origin_type) + and issubclass(origin_type, int) and field_info and hasattr(field_info, "json_schema_extra") and field_info.json_schema_extra is not None @@ -458,7 +468,7 @@ def generate_gbnf_grammar(model: type[BaseModel], processed_models: set[type[Bas if not issubclass(model, BaseModel): # For non-Pydantic classes, generate model_fields from __annotations__ or __init__ if hasattr(model, "__annotations__") and model.__annotations__: - model_fields = {name: (typ, ...) for name, typ in model.__annotations__.items()} + model_fields = {name: (typ, ...) for name, typ in get_type_hints(model).items()} else: init_signature = inspect.signature(model.__init__) parameters = init_signature.parameters @@ -466,7 +476,7 @@ def generate_gbnf_grammar(model: type[BaseModel], processed_models: set[type[Bas name != "self"} else: # For Pydantic models, use model_fields and check for ellipsis (required fields) - model_fields = model.__annotations__ + model_fields = get_type_hints(model) model_rule_parts = [] nested_rules = [] @@ -680,7 +690,7 @@ def generate_markdown_documentation( str: Generated text documentation. """ documentation = "" - pyd_models = [(model, True) for model in pydantic_models] + pyd_models: list[tuple[type[BaseModel], bool]] = [(model, True) for model in pydantic_models] for model, add_prefix in pyd_models: if add_prefix: documentation += f"{model_prefix}: {model.__name__}\n" @@ -700,9 +710,9 @@ def generate_markdown_documentation( # Indenting the fields section documentation += f" {fields_prefix}:\n" else: - documentation += f" Fields:\n" + documentation += f" Fields:\n" # noqa: F541 if isclass(model) and issubclass(model, BaseModel): - for name, field_type in model.__annotations__.items(): + for name, field_type in get_type_hints(model).items(): # if name == "markdown_code_block": # continue if get_origin(field_type) == list: @@ -750,14 +760,17 @@ def generate_field_markdown( field_info = model.model_fields.get(field_name) field_description = field_info.description if field_info and field_info.description else "" - if get_origin(field_type) == list: + origin_type = get_origin(field_type) + origin_type = field_type if origin_type is None else origin_type + + if origin_type == list: element_type = get_args(field_type)[0] field_text = f"{indent}{field_name} ({format_model_and_field_name(field_type.__name__)} of {format_model_and_field_name(element_type.__name__)})" if field_description != "": field_text += ":\n" else: field_text += "\n" - elif get_origin(field_type) == Union: + elif origin_type == Union: element_types = get_args(field_type) types = [] for element_type in element_types: @@ -778,7 +791,7 @@ def generate_field_markdown( return field_text if field_description != "": - field_text += f" Description: " + field_description + "\n" + field_text += f" Description: {field_description}\n" # Check for and include field-specific examples if available if hasattr(model, "Config") and hasattr(model.Config, @@ -788,9 +801,9 @@ def generate_field_markdown( example_text = f"'{field_example}'" if isinstance(field_example, str) else field_example field_text += f"{indent} Example: {example_text}\n" - if isclass(field_type) and issubclass(field_type, BaseModel): + if isclass(origin_type) and issubclass(origin_type, BaseModel): field_text += f"{indent} Details:\n" - for name, type_ in field_type.__annotations__.items(): + for name, type_ in get_type_hints(field_type).items(): field_text += generate_field_markdown(name, type_, field_type, depth + 2) return field_text @@ -833,7 +846,7 @@ def generate_text_documentation( str: Generated text documentation. """ documentation = "" - pyd_models = [(model, True) for model in pydantic_models] + pyd_models: list[tuple[type[BaseModel], bool]] = [(model, True) for model in pydantic_models] for model, add_prefix in pyd_models: if add_prefix: documentation += f"{model_prefix}: {model.__name__}\n" @@ -851,7 +864,7 @@ def generate_text_documentation( if isclass(model) and issubclass(model, BaseModel): documentation_fields = "" - for name, field_type in model.__annotations__.items(): + for name, field_type in get_type_hints(model).items(): # if name == "markdown_code_block": # continue if get_origin(field_type) == list: @@ -944,7 +957,7 @@ def generate_field_text( if isclass(field_type) and issubclass(field_type, BaseModel): field_text += f"{indent} Details:\n" - for name, type_ in field_type.__annotations__.items(): + for name, type_ in get_type_hints(field_type).items(): field_text += generate_field_text(name, type_, field_type, depth + 2) return field_text @@ -1164,7 +1177,7 @@ def create_dynamic_model_from_function(func: Callable[..., Any]): dynamic_fields[param.name] = ( param.annotation if param.annotation != inspect.Parameter.empty else str, default_value) # Creating the dynamic model - dynamic_model = create_model(f"{func.__name__}", **dynamic_fields) # type: ignore[call-overload] + dynamic_model = create_model(f"{func.__name__}", **dynamic_fields) for name, param_doc in param_docs: dynamic_model.model_fields[name].description = param_doc.description @@ -1228,9 +1241,6 @@ def map_grammar_names_to_pydantic_model_class(pydantic_model_list): return output -from enum import Enum - - def json_schema_to_python_types(schema): type_map = { "any": Any, @@ -1275,7 +1285,7 @@ def convert_dictionary_to_pydantic_model(dictionary: dict[str, Any], model_name: if items != {}: array = {"properties": items} array_type = convert_dictionary_to_pydantic_model(array, f"{model_name}_{field_name}_items") - fields[field_name] = (List[array_type], ...) # type: ignore[valid-type] + fields[field_name] = (List[array_type], ...) else: fields[field_name] = (list, ...) elif field_type == "object": @@ -1285,7 +1295,8 @@ def convert_dictionary_to_pydantic_model(dictionary: dict[str, Any], model_name: required = field_data.get("enum", []) for key, field in fields.items(): if key not in required: - fields[key] = (Optional[fields[key][0]], ...) + optional_type = fields[key][0] + fields[key] = (Optional[optional_type], ...) else: field_type = json_schema_to_python_types(field_type) fields[field_name] = (field_type, ...) @@ -1305,6 +1316,7 @@ def convert_dictionary_to_pydantic_model(dictionary: dict[str, Any], model_name: required = dictionary.get("required", []) for key, field in fields.items(): if key not in required: - fields[key] = (Optional[fields[key][0]], ...) + optional_type = fields[key][0] + fields[key] = (Optional[optional_type], ...) custom_model = create_model(model_name, **fields) return custom_model diff --git a/examples/pydantic_models_to_grammar_examples.py b/examples/pydantic_models_to_grammar_examples.py new file mode 100755 index 000000000..eb000d5cc --- /dev/null +++ b/examples/pydantic_models_to_grammar_examples.py @@ -0,0 +1,312 @@ +#!/usr/bin/env python3 + +"""Function calling example using pydantic models.""" + +from __future__ import annotations + +import argparse +import datetime +import json +import logging +import textwrap +import sys +from enum import Enum +from typing import Optional, Union + +import requests +from pydantic import BaseModel, Field +from pydantic_models_to_grammar import (add_run_method_to_dynamic_model, convert_dictionary_to_pydantic_model, + create_dynamic_model_from_function, generate_gbnf_grammar_and_documentation) + + +def create_completion(host, prompt, gbnf_grammar): + """Calls the /completion API on llama-server. + + See + https://github.com/ggerganov/llama.cpp/tree/HEAD/examples/server#api-endpoints + """ + print(f" Request:\n Grammar:\n{textwrap.indent(gbnf_grammar, ' ')}\n Prompt:\n{textwrap.indent(prompt.rstrip(), ' ')}") + headers = {"Content-Type": "application/json"} + data = {"prompt": prompt, "grammar": gbnf_grammar} + result = requests.post(f"http://{host}/completion", headers=headers, json=data).json() + assert data.get("error") is None, data + logging.info("Result: %s", result) + content = result["content"] + print(f" Model: {result['model']}") + print(f" Result:\n{textwrap.indent(json.dumps(json.loads(content), indent=2), ' ')}") + return content + + +# A function for the agent to send a message to the user. +class SendMessageToUser(BaseModel): + """Send a message to the User.""" + chain_of_thought: str = Field(..., description="Your chain of thought while sending the message.") + message: str = Field(..., description="Message you want to send to the user.") + + def run(self): + print(f"SendMessageToUser: {self.message}") + + +def example_rce(host): + """Minimal test case where the LLM call an arbitrary python function.""" + print("- example_rce") + tools = [SendMessageToUser] + gbnf_grammar, documentation = generate_gbnf_grammar_and_documentation( + pydantic_model_list=tools, outer_object_name="function", + outer_object_content="function_parameters", model_prefix="Function", fields_prefix="Parameters") + system_message = "You are an advanced AI, tasked to assist the user by calling functions in JSON format. The following are the available functions and their parameters and types:\n\n" + documentation + user_message = "What is 42 * 42?" + prompt = f"<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{user_message}<|im_end|>\n<|im_start|>assistant" + text = create_completion(host, prompt, gbnf_grammar) + json_data = json.loads(text) + tools_map = {tool.__name__:tool for tool in tools} + # This finds "SendMessageToUser": + tool = tools_map.get(json_data["function"]) + if not tool: + print(f"Error: unknown tool {json_data['function']}") + return 1 + tool(**json_data["function_parameters"]).run() + return 0 + + +# Enum for the calculator tool. +class MathOperation(Enum): + ADD = "add" + SUBTRACT = "subtract" + MULTIPLY = "multiply" + DIVIDE = "divide" + + +# Simple pydantic calculator tool for the agent that can add, subtract, +# multiply, and divide. Docstring and description of fields will be used in +# system prompt. +class Calculator(BaseModel): + """Perform a math operation on two numbers.""" + number_one: Union[int, float] = Field(..., description="First number.") + operation: MathOperation = Field(..., description="Math operation to perform.") + number_two: Union[int, float] = Field(..., description="Second number.") + + def run(self): + if self.operation == MathOperation.ADD: + return self.number_one + self.number_two + elif self.operation == MathOperation.SUBTRACT: + return self.number_one - self.number_two + elif self.operation == MathOperation.MULTIPLY: + return self.number_one * self.number_two + elif self.operation == MathOperation.DIVIDE: + return self.number_one / self.number_two + else: + raise ValueError("Unknown operation.") + + +def example_calculator(host): + """Have the LLM ask to get a calculation done. + + Here the grammar gets generated by passing the available function models to + generate_gbnf_grammar_and_documentation function. This also generates a + documentation usable by the LLM. + + pydantic_model_list is the list of pydantic models outer_object_name is an + optional name for an outer object around the actual model object. Like a + "function" object with "function_parameters" which contains the actual model + object. If None, no outer object will be generated outer_object_content is + the name of outer object content. + + model_prefix is the optional prefix for models in the documentation. (Default="Output Model") + fields_prefix is the prefix for the model fields in the documentation. (Default="Output Fields") + """ + print("- example_calculator") + tools = [SendMessageToUser, Calculator] + gbnf_grammar, documentation = generate_gbnf_grammar_and_documentation( + pydantic_model_list=tools, outer_object_name="function", + outer_object_content="function_parameters", model_prefix="Function", fields_prefix="Parameters") + system_message = "You are an advanced AI, tasked to assist the user by calling functions in JSON format. The following are the available functions and their parameters and types:\n\n" + documentation + user_message1 = "What is 42 * 42?" + prompt = f"<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{user_message1}<|im_end|>\n<|im_start|>assistant" + text = create_completion(host, prompt, gbnf_grammar) + json_data = json.loads(text) + expected = { + "function": "Calculator", + "function_parameters": { + "number_one": 42, + "operation": "multiply", + "number_two": 42 + } + } + if json_data != expected: + print(" Result is not as expected!") + tools_map = {tool.__name__:tool for tool in tools} + # This finds "Calculator": + tool = tools_map.get(json_data["function"]) + if not tool: + print(f"Error: unknown tool {json_data['function']}") + return 1 + result = tool(**json_data["function_parameters"]).run() + print(f" Call {json_data['function']} gave result {result}") + return 0 + + +class Category(Enum): + """The category of the book.""" + Fiction = "Fiction" + NonFiction = "Non-Fiction" + + +class Book(BaseModel): + """Represents an entry about a book.""" + title: str = Field(..., description="Title of the book.") + author: str = Field(..., description="Author of the book.") + published_year: Optional[int] = Field(..., description="Publishing year of the book.") + keywords: list[str] = Field(..., description="A list of keywords.") + category: Category = Field(..., description="Category of the book.") + summary: str = Field(..., description="Summary of the book.") + + +def example_struct(host): + """A example structured output based on pydantic models. + + The LLM will create an entry for a Book database out of an unstructured + text. We need no additional parameters other than our list of pydantic + models. + """ + print("- example_struct") + tools = [Book] + gbnf_grammar, documentation = generate_gbnf_grammar_and_documentation(pydantic_model_list=tools) + system_message = "You are an advanced AI, tasked to create a dataset entry in JSON for a Book. The following is the expected output model:\n\n" + documentation + text = """The Feynman Lectures on Physics is a physics textbook based on some lectures by Richard Feynman, a Nobel laureate who has sometimes been called "The Great Explainer". The lectures were presented before undergraduate students at the California Institute of Technology (Caltech), during 1961–1963. The book's co-authors are Feynman, Robert B. Leighton, and Matthew Sands.""" + prompt = f"<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{text}<|im_end|>\n<|im_start|>assistant" + text = create_completion(host, prompt, gbnf_grammar) + json_data = json.loads(text) + # In this case, there's no function nor function_parameters. + # Here the result will vary based on the LLM used. + keys = sorted(["title", "author", "published_year", "keywords", "category", "summary"]) + if keys != sorted(json_data.keys()): + print(f"Unexpected result: {sorted(json_data.keys())}") + return 1 + book = Book(**json_data) + print(f" As a Book object: %s" % book) + return 0 + + +def get_current_datetime(output_format: Optional[str] = None): + """Get the current date and time in the given format. + + Args: + output_format: formatting string for the date and time, defaults to '%Y-%m-%d %H:%M:%S' + """ + return datetime.datetime.now().strftime(output_format or "%Y-%m-%d %H:%M:%S") + + +# Example function to get the weather. +def get_current_weather(location, unit): + """Get the current weather in a given location""" + if "London" in location: + return json.dumps({"location": "London", "temperature": "42", "unit": unit.value}) + elif "New York" in location: + return json.dumps({"location": "New York", "temperature": "24", "unit": unit.value}) + elif "North Pole" in location: + return json.dumps({"location": "North Pole", "temperature": "-42", "unit": unit.value}) + return json.dumps({"location": location, "temperature": "unknown"}) + + +def example_concurrent(host): + """An example for parallel function calling with a Python function, a pydantic + function model and an OpenAI like function definition. + """ + print("- example_concurrent") + # Function definition in OpenAI style. + current_weather_tool = { + "type": "function", + "function": { + "name": "get_current_weather", + "description": "Get the current weather in a given location", + "parameters": { + "type": "object", + "properties": { + "location": { + "type": "string", + "description": "The city and state, e.g. San Francisco, CA", + }, + "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]}, + }, + "required": ["location"], + }, + }, + } + # Convert OpenAI function definition into pydantic model. + current_weather_tool_model = convert_dictionary_to_pydantic_model(current_weather_tool) + # Add the actual function to a pydantic model. + current_weather_tool_model = add_run_method_to_dynamic_model(current_weather_tool_model, get_current_weather) + + # Convert normal Python function to a pydantic model. + current_datetime_model = create_dynamic_model_from_function(get_current_datetime) + + tools = [SendMessageToUser, Calculator, current_datetime_model, current_weather_tool_model] + gbnf_grammar, documentation = generate_gbnf_grammar_and_documentation( + pydantic_model_list=tools, outer_object_name="function", + outer_object_content="params", model_prefix="Function", fields_prefix="Parameters", list_of_outputs=True) + system_message = "You are an advanced AI assistant. You are interacting with the user and with your environment by calling functions. You call functions by writing JSON objects, which represent specific function calls.\nBelow is a list of your available function calls:\n\n" + documentation + text = """Get the date and time, get the current weather in celsius in London and solve the following calculation: 42 * 42""" + prompt = f"<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{text}<|im_end|>\n<|im_start|>assistant" + text = create_completion(host, prompt, gbnf_grammar) + json_data = json.loads(text) + expected = [ + { + "function": "get_current_datetime", + "params": { + "output_format": "%Y-%m-%d %H:%M:%S" + } + }, + { + "function": "get_current_weather", + "params": { + "location": "London", + "unit": "celsius" + } + }, + { + "function": "Calculator", + "params": { + "number_one": 42, + "operation": "multiply", + "number_two": 42 + } + } + ] + res = 0 + if json_data != expected: + print(" Result is not as expected!") + print(" This can happen on highly quantized models") + res = 1 + tools_map = {tool.__name__:tool for tool in tools} + for call in json_data: + tool = tools_map.get(call["function"]) + if not tool: + print(f"Error: unknown tool {call['function']}") + return 1 + result = tool(**call["params"]).run() + print(f" Call {call['function']} returned {result}") + # Should output something like this: + # Call get_current_datetime returned 2024-07-15 09:50:38 + # Call get_current_weather returned {"location": "London", "temperature": "42", "unit": "celsius"} + # Call Calculator returned 1764 + return res + + +def main(): + parser = argparse.ArgumentParser(description=sys.modules[__name__].__doc__) + parser.add_argument("--host", default="localhost:8080", help="llama.cpp server") + parser.add_argument("-v", "--verbose", action="store_true", help="enables logging") + args = parser.parse_args() + logging.basicConfig(level=logging.INFO if args.verbose else logging.ERROR) + ret = 0 + # Comment out below to only run the example you want. + ret = ret or example_rce(args.host) + ret = ret or example_calculator(args.host) + ret = ret or example_struct(args.host) + ret = ret or example_concurrent(args.host) + return ret + + +if __name__ == "__main__": + sys.exit(main()) diff --git a/examples/quantize-stats/quantize-stats.cpp b/examples/quantize-stats/quantize-stats.cpp index 746df8446..68cf8d359 100644 --- a/examples/quantize-stats/quantize-stats.cpp +++ b/examples/quantize-stats/quantize-stats.cpp @@ -154,7 +154,7 @@ static void test_roundtrip_on_chunk( } if (use_reference) { - qfns.from_float_reference(input_scratch, quantized_scratch, chunk_size); + qfns.from_float_ref(input_scratch, quantized_scratch, chunk_size); } else { qfns.from_float(input_scratch, quantized_scratch, chunk_size); } diff --git a/examples/quantize/README.md b/examples/quantize/README.md index b78ece4e7..553c2701b 100644 --- a/examples/quantize/README.md +++ b/examples/quantize/README.md @@ -4,7 +4,89 @@ You can also use the [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf- Note: It is synced from llama.cpp `main` every 6 hours. -## Llama 2 7B +Example usage: + +```bash +# obtain the official LLaMA model weights and place them in ./models +ls ./models +llama-2-7b tokenizer_checklist.chk tokenizer.model +# [Optional] for models using BPE tokenizers +ls ./models + vocab.json +# [Optional] for PyTorch .bin models like Mistral-7B +ls ./models + + +# install Python dependencies +python3 -m pip install -r requirements.txt + +# convert the model to ggml FP16 format +python3 convert_hf_to_gguf.py models/mymodel/ + +# quantize the model to 4-bits (using Q4_K_M method) +./llama-quantize ./models/mymodel/ggml-model-f16.gguf ./models/mymodel/ggml-model-Q4_K_M.gguf Q4_K_M + +# update the gguf filetype to current version if older version is now unsupported +./llama-quantize ./models/mymodel/ggml-model-Q4_K_M.gguf ./models/mymodel/ggml-model-Q4_K_M-v2.gguf COPY +``` + +Run the quantized model: + +```bash +# start inference on a gguf model +./llama-cli -m ./models/mymodel/ggml-model-Q4_K_M.gguf -n 128 +``` + +When running the larger models, make sure you have enough disk space to store all the intermediate files. + +## Memory/Disk Requirements + +As the models are currently fully loaded into memory, you will need adequate disk space to save them and sufficient RAM to load them. At the moment, memory and disk requirements are the same. + +| Model | Original size | Quantized size (Q4_0) | +|------:|--------------:|----------------------:| +| 7B | 13 GB | 3.9 GB | +| 13B | 24 GB | 7.8 GB | +| 30B | 60 GB | 19.5 GB | +| 65B | 120 GB | 38.5 GB | + +## Quantization + +Several quantization methods are supported. They differ in the resulting model disk size and inference speed. + +*(outdated)* + +| Model | Measure | F16 | Q4_0 | Q4_1 | Q5_0 | Q5_1 | Q8_0 | +|------:|--------------|-------:|-------:|-------:|-------:|-------:|-------:| +| 7B | perplexity | 5.9066 | 6.1565 | 6.0912 | 5.9862 | 5.9481 | 5.9070 | +| 7B | file size | 13.0G | 3.5G | 3.9G | 4.3G | 4.7G | 6.7G | +| 7B | ms/tok @ 4th | 127 | 55 | 54 | 76 | 83 | 72 | +| 7B | ms/tok @ 8th | 122 | 43 | 45 | 52 | 56 | 67 | +| 7B | bits/weight | 16.0 | 4.5 | 5.0 | 5.5 | 6.0 | 8.5 | +| 13B | perplexity | 5.2543 | 5.3860 | 5.3608 | 5.2856 | 5.2706 | 5.2548 | +| 13B | file size | 25.0G | 6.8G | 7.6G | 8.3G | 9.1G | 13G | +| 13B | ms/tok @ 4th | - | 103 | 105 | 148 | 160 | 131 | +| 13B | ms/tok @ 8th | - | 73 | 82 | 98 | 105 | 128 | +| 13B | bits/weight | 16.0 | 4.5 | 5.0 | 5.5 | 6.0 | 8.5 | + +- [k-quants](https://github.com/ggerganov/llama.cpp/pull/1684) +- recent k-quants improvements and new i-quants + - [#2707](https://github.com/ggerganov/llama.cpp/pull/2707) + - [#2807](https://github.com/ggerganov/llama.cpp/pull/2807) + - [#4773 - 2-bit i-quants (inference)](https://github.com/ggerganov/llama.cpp/pull/4773) + - [#4856 - 2-bit i-quants (inference)](https://github.com/ggerganov/llama.cpp/pull/4856) + - [#4861 - importance matrix](https://github.com/ggerganov/llama.cpp/pull/4861) + - [#4872 - MoE models](https://github.com/ggerganov/llama.cpp/pull/4872) + - [#4897 - 2-bit quantization](https://github.com/ggerganov/llama.cpp/pull/4897) + - [#4930 - imatrix for all k-quants](https://github.com/ggerganov/llama.cpp/pull/4930) + - [#4951 - imatrix on the GPU](https://github.com/ggerganov/llama.cpp/pull/4957) + - [#4969 - imatrix for legacy quants](https://github.com/ggerganov/llama.cpp/pull/4969) + - [#4996 - k-qunats tuning](https://github.com/ggerganov/llama.cpp/pull/4996) + - [#5060 - Q3_K_XS](https://github.com/ggerganov/llama.cpp/pull/5060) + - [#5196 - 3-bit i-quants](https://github.com/ggerganov/llama.cpp/pull/5196) + - [quantization tuning](https://github.com/ggerganov/llama.cpp/pull/5320), [another one](https://github.com/ggerganov/llama.cpp/pull/5334), and [another one](https://github.com/ggerganov/llama.cpp/pull/5361) + +**Llama 2 7B** | Quantization | Bits per Weight (BPW) | |--------------|-----------------------| @@ -18,7 +100,8 @@ Note: It is synced from llama.cpp `main` every 6 hours. | Q5_K_M | 5.68 | | Q6_K | 6.56 | -## Llama 2 13B +**Llama 2 13B** + Quantization | Bits per Weight (BPW) -- | -- Q2_K | 3.34 @@ -31,7 +114,7 @@ Q5_K_S | 5.51 Q5_K_M | 5.67 Q6_K | 6.56 -# Llama 2 70B +**Llama 2 70B** Quantization | Bits per Weight (BPW) -- | -- diff --git a/examples/quantize/quantize.cpp b/examples/quantize/quantize.cpp index aed39a4d0..88044e622 100644 --- a/examples/quantize/quantize.cpp +++ b/examples/quantize/quantize.cpp @@ -16,43 +16,46 @@ struct quant_option { }; static const std::vector QUANT_OPTIONS = { - { "Q4_0", LLAMA_FTYPE_MOSTLY_Q4_0, " 4.34G, +0.4685 ppl @ Llama-3-8B", }, - { "Q4_1", LLAMA_FTYPE_MOSTLY_Q4_1, " 4.78G, +0.4511 ppl @ Llama-3-8B", }, - { "Q5_0", LLAMA_FTYPE_MOSTLY_Q5_0, " 5.21G, +0.1316 ppl @ Llama-3-8B", }, - { "Q5_1", LLAMA_FTYPE_MOSTLY_Q5_1, " 5.65G, +0.1062 ppl @ Llama-3-8B", }, - { "IQ2_XXS",LLAMA_FTYPE_MOSTLY_IQ2_XXS," 2.06 bpw quantization", }, - { "IQ2_XS", LLAMA_FTYPE_MOSTLY_IQ2_XS, " 2.31 bpw quantization", }, - { "IQ2_S", LLAMA_FTYPE_MOSTLY_IQ2_S, " 2.5 bpw quantization", }, - { "IQ2_M", LLAMA_FTYPE_MOSTLY_IQ2_M, " 2.7 bpw quantization", }, - { "IQ1_S", LLAMA_FTYPE_MOSTLY_IQ1_S, " 1.56 bpw quantization", }, - { "IQ1_M", LLAMA_FTYPE_MOSTLY_IQ1_M, " 1.75 bpw quantization", }, - { "Q1_3", LLAMA_FTYPE_MOSTLY_Q1_3, " 1.63 bpw for BitNet b1.58", }, - { "Q2_2", LLAMA_FTYPE_MOSTLY_Q2_2, " 2.00 bpw for BitNet b1.58", }, - { "Q2_K", LLAMA_FTYPE_MOSTLY_Q2_K, " 2.96G, +3.5199 ppl @ Llama-3-8B", }, - { "Q2_K_S", LLAMA_FTYPE_MOSTLY_Q2_K_S, " 2.96G, +3.1836 ppl @ Llama-3-8B", }, - { "IQ3_XXS",LLAMA_FTYPE_MOSTLY_IQ3_XXS," 3.06 bpw quantization", }, - { "IQ3_S", LLAMA_FTYPE_MOSTLY_IQ3_S, " 3.44 bpw quantization", }, - { "IQ3_M", LLAMA_FTYPE_MOSTLY_IQ3_M, " 3.66 bpw quantization mix", }, - { "Q3_K", LLAMA_FTYPE_MOSTLY_Q3_K_M, "alias for Q3_K_M" }, - { "IQ3_XS", LLAMA_FTYPE_MOSTLY_IQ3_XS, " 3.3 bpw quantization", }, - { "Q3_K_S", LLAMA_FTYPE_MOSTLY_Q3_K_S, " 3.41G, +1.6321 ppl @ Llama-3-8B", }, - { "Q3_K_M", LLAMA_FTYPE_MOSTLY_Q3_K_M, " 3.74G, +0.6569 ppl @ Llama-3-8B", }, - { "Q3_K_L", LLAMA_FTYPE_MOSTLY_Q3_K_L, " 4.03G, +0.5562 ppl @ Llama-3-8B", }, - { "IQ4_NL", LLAMA_FTYPE_MOSTLY_IQ4_NL, " 4.50 bpw non-linear quantization", }, - { "IQ4_XS", LLAMA_FTYPE_MOSTLY_IQ4_XS, " 4.25 bpw non-linear quantization", }, - { "Q4_K", LLAMA_FTYPE_MOSTLY_Q4_K_M, "alias for Q4_K_M", }, - { "Q4_K_S", LLAMA_FTYPE_MOSTLY_Q4_K_S, " 4.37G, +0.2689 ppl @ Llama-3-8B", }, - { "Q4_K_M", LLAMA_FTYPE_MOSTLY_Q4_K_M, " 4.58G, +0.1754 ppl @ Llama-3-8B", }, - { "Q5_K", LLAMA_FTYPE_MOSTLY_Q5_K_M, "alias for Q5_K_M", }, - { "Q5_K_S", LLAMA_FTYPE_MOSTLY_Q5_K_S, " 5.21G, +0.1049 ppl @ Llama-3-8B", }, - { "Q5_K_M", LLAMA_FTYPE_MOSTLY_Q5_K_M, " 5.33G, +0.0569 ppl @ Llama-3-8B", }, - { "Q6_K", LLAMA_FTYPE_MOSTLY_Q6_K, " 6.14G, +0.0217 ppl @ Llama-3-8B", }, - { "Q8_0", LLAMA_FTYPE_MOSTLY_Q8_0, " 7.96G, +0.0026 ppl @ Llama-3-8B", }, - { "F16", LLAMA_FTYPE_MOSTLY_F16, "14.00G, +0.0020 ppl @ Mistral-7B", }, - { "BF16", LLAMA_FTYPE_MOSTLY_BF16, "14.00G, -0.0050 ppl @ Mistral-7B", }, - { "F32", LLAMA_FTYPE_ALL_F32, "26.00G @ 7B", }, + { "Q4_0", LLAMA_FTYPE_MOSTLY_Q4_0, " 4.34G, +0.4685 ppl @ Llama-3-8B", }, + { "Q4_1", LLAMA_FTYPE_MOSTLY_Q4_1, " 4.78G, +0.4511 ppl @ Llama-3-8B", }, + { "Q5_0", LLAMA_FTYPE_MOSTLY_Q5_0, " 5.21G, +0.1316 ppl @ Llama-3-8B", }, + { "Q5_1", LLAMA_FTYPE_MOSTLY_Q5_1, " 5.65G, +0.1062 ppl @ Llama-3-8B", }, + { "IQ2_XXS", LLAMA_FTYPE_MOSTLY_IQ2_XXS, " 2.06 bpw quantization", }, + { "IQ2_XS", LLAMA_FTYPE_MOSTLY_IQ2_XS, " 2.31 bpw quantization", }, + { "IQ2_S", LLAMA_FTYPE_MOSTLY_IQ2_S, " 2.5 bpw quantization", }, + { "IQ2_M", LLAMA_FTYPE_MOSTLY_IQ2_M, " 2.7 bpw quantization", }, + { "IQ1_S", LLAMA_FTYPE_MOSTLY_IQ1_S, " 1.56 bpw quantization", }, + { "IQ1_M", LLAMA_FTYPE_MOSTLY_IQ1_M, " 1.75 bpw quantization", }, + { "Q1_3", LLAMA_FTYPE_MOSTLY_Q1_3, " 1.63 bpw for BitNet b1.58", }, + { "Q2_2", LLAMA_FTYPE_MOSTLY_Q2_2, " 2.00 bpw for BitNet b1.58", }, + { "Q2_K", LLAMA_FTYPE_MOSTLY_Q2_K, " 2.96G, +3.5199 ppl @ Llama-3-8B", }, + { "Q2_K_S", LLAMA_FTYPE_MOSTLY_Q2_K_S, " 2.96G, +3.1836 ppl @ Llama-3-8B", }, + { "IQ3_XXS", LLAMA_FTYPE_MOSTLY_IQ3_XXS, " 3.06 bpw quantization", }, + { "IQ3_S", LLAMA_FTYPE_MOSTLY_IQ3_S, " 3.44 bpw quantization", }, + { "IQ3_M", LLAMA_FTYPE_MOSTLY_IQ3_M, " 3.66 bpw quantization mix", }, + { "Q3_K", LLAMA_FTYPE_MOSTLY_Q3_K_M, "alias for Q3_K_M" }, + { "IQ3_XS", LLAMA_FTYPE_MOSTLY_IQ3_XS, " 3.3 bpw quantization", }, + { "Q3_K_S", LLAMA_FTYPE_MOSTLY_Q3_K_S, " 3.41G, +1.6321 ppl @ Llama-3-8B", }, + { "Q3_K_M", LLAMA_FTYPE_MOSTLY_Q3_K_M, " 3.74G, +0.6569 ppl @ Llama-3-8B", }, + { "Q3_K_L", LLAMA_FTYPE_MOSTLY_Q3_K_L, " 4.03G, +0.5562 ppl @ Llama-3-8B", }, + { "IQ4_NL", LLAMA_FTYPE_MOSTLY_IQ4_NL, " 4.50 bpw non-linear quantization", }, + { "IQ4_XS", LLAMA_FTYPE_MOSTLY_IQ4_XS, " 4.25 bpw non-linear quantization", }, + { "Q4_K", LLAMA_FTYPE_MOSTLY_Q4_K_M, "alias for Q4_K_M", }, + { "Q4_K_S", LLAMA_FTYPE_MOSTLY_Q4_K_S, " 4.37G, +0.2689 ppl @ Llama-3-8B", }, + { "Q4_K_M", LLAMA_FTYPE_MOSTLY_Q4_K_M, " 4.58G, +0.1754 ppl @ Llama-3-8B", }, + { "Q5_K", LLAMA_FTYPE_MOSTLY_Q5_K_M, "alias for Q5_K_M", }, + { "Q5_K_S", LLAMA_FTYPE_MOSTLY_Q5_K_S, " 5.21G, +0.1049 ppl @ Llama-3-8B", }, + { "Q5_K_M", LLAMA_FTYPE_MOSTLY_Q5_K_M, " 5.33G, +0.0569 ppl @ Llama-3-8B", }, + { "Q6_K", LLAMA_FTYPE_MOSTLY_Q6_K, " 6.14G, +0.0217 ppl @ Llama-3-8B", }, + { "Q8_0", LLAMA_FTYPE_MOSTLY_Q8_0, " 7.96G, +0.0026 ppl @ Llama-3-8B", }, + { "Q4_0_4_4", LLAMA_FTYPE_MOSTLY_Q4_0_4_4, " 4.34G, +0.4685 ppl @ Llama-3-8B", }, + { "Q4_0_4_8", LLAMA_FTYPE_MOSTLY_Q4_0_4_8, " 4.34G, +0.4685 ppl @ Llama-3-8B", }, + { "Q4_0_8_8", LLAMA_FTYPE_MOSTLY_Q4_0_8_8, " 4.34G, +0.4685 ppl @ Llama-3-8B", }, + { "F16", LLAMA_FTYPE_MOSTLY_F16, "14.00G, +0.0020 ppl @ Mistral-7B", }, + { "BF16", LLAMA_FTYPE_MOSTLY_BF16, "14.00G, -0.0050 ppl @ Mistral-7B", }, + { "F32", LLAMA_FTYPE_ALL_F32, "26.00G @ 7B", }, // Note: Ensure COPY comes after F32 to avoid ftype 0 from matching. - { "COPY", LLAMA_FTYPE_ALL_F32, "only copy tensors, no quantizing", }, + { "COPY", LLAMA_FTYPE_ALL_F32, "only copy tensors, no quantizing", }, }; static const char * const LLM_KV_QUANTIZE_IMATRIX_FILE = "quantize.imatrix.file"; diff --git a/examples/regex-to-grammar.py b/examples/regex_to_grammar.py similarity index 100% rename from examples/regex-to-grammar.py rename to examples/regex_to_grammar.py diff --git a/examples/save-load-state/save-load-state.cpp b/examples/save-load-state/save-load-state.cpp index 00c2277ac..d8afdc141 100644 --- a/examples/save-load-state/save-load-state.cpp +++ b/examples/save-load-state/save-load-state.cpp @@ -47,7 +47,7 @@ int main(int argc, char ** argv) { // save state (rng, logits, embedding and kv_cache) to file { std::vector state_mem(llama_state_get_size(ctx)); - const size_t written = llama_state_get_data(ctx, state_mem.data()); + const size_t written = llama_state_get_data(ctx, state_mem.data(), state_mem.size()); FILE *fp_write = fopen("dump_state.bin", "wb"); fwrite(state_mem.data(), 1, written, fp_write); @@ -99,13 +99,16 @@ int main(int argc, char ** argv) { // load state (rng, logits, embedding and kv_cache) from file { - std::vector state_mem(llama_state_get_size(ctx2)); + std::vector state_mem; FILE * fp_read = fopen("dump_state.bin", "rb"); + fseek(fp_read, 0, SEEK_END); + state_mem.resize(ftell(fp_read)); + fseek(fp_read, 0, SEEK_SET); const size_t read = fread(state_mem.data(), 1, state_mem.size(), fp_read); fclose(fp_read); - if (read != llama_state_set_data(ctx2, state_mem.data())) { + if (read != llama_state_set_data(ctx2, state_mem.data(), state_mem.size())) { fprintf(stderr, "\n%s : failed to read state\n", __func__); llama_free(ctx2); llama_free_model(model); @@ -159,13 +162,16 @@ int main(int argc, char ** argv) { // load state (rng, logits, embedding and kv_cache) from file { - std::vector state_mem(llama_state_get_size(ctx3)); + std::vector state_mem; FILE * fp_read = fopen("dump_state.bin", "rb"); + fseek(fp_read, 0, SEEK_END); + state_mem.resize(ftell(fp_read)); + fseek(fp_read, 0, SEEK_SET); const size_t read = fread(state_mem.data(), 1, state_mem.size(), fp_read); fclose(fp_read); - if (read != llama_state_set_data(ctx3, state_mem.data())) { + if (read != llama_state_set_data(ctx3, state_mem.data(), state_mem.size())) { fprintf(stderr, "\n%s : failed to read state\n", __func__); llama_free(ctx3); llama_free_model(model); @@ -182,7 +188,7 @@ int main(int argc, char ** argv) { { // save kv of seq 0 std::vector seq_store(llama_state_seq_get_size(ctx3, 0)); - const size_t ncopy = llama_state_seq_get_data(ctx3, seq_store.data(), 0); + const size_t ncopy = llama_state_seq_get_data(ctx3, seq_store.data(), seq_store.size(), 0); if (ncopy != seq_store.size()) { fprintf(stderr, "\n%s : seq copy data length %zd does not match expected length %zd\n", __func__, ncopy, seq_store.size()); llama_free(ctx3); @@ -196,7 +202,7 @@ int main(int argc, char ** argv) { fprintf(stderr, "%s : kv cache cleared\n", __func__); // restore kv into seq 1 - const size_t nset = llama_state_seq_set_data(ctx3, seq_store.data(), 1); + const size_t nset = llama_state_seq_set_data(ctx3, seq_store.data(), seq_store.size(), 1); if (nset != seq_store.size()) { fprintf(stderr, "\n%s : seq set data length %zd does not match expected length %zd\n", __func__, nset, seq_store.size()); llama_free(ctx3); diff --git a/examples/server/README.md b/examples/server/README.md index e7fb0bf64..33a2b95cc 100644 --- a/examples/server/README.md +++ b/examples/server/README.md @@ -5,7 +5,7 @@ Fast, lightweight, pure C/C++ HTTP server based on [httplib](https://github.com/ Set of LLM REST APIs and a simple web front end to interact with llama.cpp. **Features:** - * LLM inference of F16 and quantum models on GPU and CPU + * LLM inference of F16 and quantized models on GPU and CPU * [OpenAI API](https://github.com/openai/openai-openapi) compatible chat completions and embeddings routes * Parallel decoding with multi-user support * Continuous batching @@ -15,68 +15,281 @@ Set of LLM REST APIs and a simple web front end to interact with llama.cpp. The project is under active development, and we are [looking for feedback and contributors](https://github.com/ggerganov/llama.cpp/issues/4216). -**Command line options:** +## Usage -- `-v`, `--verbose`: Enable verbose server output. When using the `/completion` endpoint, this includes the tokenized prompt, the full request and the full response. -- `-t N`, `--threads N`: Set the number of threads to use by CPU layers during generation. Not used by model layers that are offloaded to GPU. This option has no effect when using the maximum number of GPU layers. Default: `std::thread::hardware_concurrency()` (number of CPU cores). -- `-tb N, --threads-batch N`: Set the number of threads to use by CPU layers during batch and prompt processing (>= 32 tokens). This option has no effect if a GPU is available. Default: `--threads`. -- `--threads-http N`: Number of threads in the http server pool to process requests. Default: `max(std::thread::hardware_concurrency() - 1, --parallel N + 2)` -- `-m FNAME`, `--model FNAME`: Specify the path to the LLaMA model file (e.g., `models/7B/ggml-model.gguf`). -- `-mu MODEL_URL --model-url MODEL_URL`: Specify a remote http url to download the file. Default: unused -- `-hfr REPO, --hf-repo REPO`: Hugging Face model repository. Default: unused -- `-hff FILE, --hf-file FILE`: Hugging Face model file. Default: unused -- `-a ALIAS`, `--alias ALIAS`: Set an alias for the model. The alias will be returned in API responses. -- `-c N`, `--ctx-size N`: Set the size of the prompt context. The default is `512`, but LLaMA models were built with a context of `2048`, which will provide better results for longer input/inference. The size may differ in other models, for example, baichuan models were build with a context of `4096`. -- `-ngl N`, `--n-gpu-layers N`: When compiled with GPU support, this option allows offloading some layers to the GPU for computation. Generally results in increased performance. -- `-mg i, --main-gpu i`: When using multiple GPUs, this option controls which GPU is used for small tensors for which the overhead of splitting the computation across all GPUs is not worthwhile. The GPU in question will use slightly more VRAM to store a scratch buffer for temporary results. By default, GPU `0` is used. -- `-ts SPLIT, --tensor-split SPLIT`: When using multiple GPUs, this option controls how large tensors should be split across all GPUs. `SPLIT` is a comma-separated list of non-negative values that assigns the proportion of data that each GPU should get in order. For example, "3,2" will assign 60% of the data to GPU 0 and 40% to GPU 1. By default, the data is split in proportion to VRAM, but this may not be optimal for performance. -- `-b N`, `--batch-size N`: Set the batch size for prompt processing. Default: `2048` -- `-ub N`, `--ubatch-size N`: Physical maximum batch size. Default: `512` -- `--mlock`: Lock the model in memory, preventing it from being swapped out when memory-mapped. -- `--no-mmap`: Do not memory-map the model. By default, models are mapped into memory, which allows the system to load only the necessary parts of the model as needed. -- `--numa STRATEGY`: Attempt one of the below optimization strategies that may help on some NUMA systems -- `--numa distribute`: Spread execution evenly over all nodes -- `--numa isolate`: Only spawn threads on CPUs on the node that execution started on -- `--numa numactl`: Use the CPU map provided by numactl. If run without this previously, it is recommended to drop the system page cache before using this. See https://github.com/ggerganov/llama.cpp/issues/1437 -- `--numa`: Attempt optimizations that may help on some NUMA systems. -- `--lora FNAME`: Apply a LoRA (Low-Rank Adaptation) adapter to the model (implies --no-mmap). This allows you to adapt the pretrained model to specific tasks or domains. -- `--lora-base FNAME`: Optional model to use as a base for the layers modified by the LoRA adapter. This flag is used in conjunction with the `--lora` flag, and specifies the base model for the adaptation. -- `-to N`, `--timeout N`: Server read/write timeout in seconds. Default `600` -- `--host`: Set the hostname or ip address to listen. Default `127.0.0.1` -- `--port`: Set the port to listen. Default: `8080` -- `--path`: Path from which to serve static files. Default: disabled -- `--api-key`: Set an api key for request authorization. By default, the server responds to every request. With an api key set, the requests must have the Authorization header set with the api key as Bearer token. May be used multiple times to enable multiple valid keys. -- `--api-key-file`: Path to file containing api keys delimited by new lines. If set, requests must include one of the keys for access. May be used in conjunction with `--api-key`s. -- `--embeddings`: Enable embedding vector output and the OAI compatible endpoint /v1/embeddings. Physical batch size (`--ubatch-size`) must be carefully defined. Default: disabled -- `-np N`, `--parallel N`: Set the number of slots for process requests. Default: `1`. Values > 1 will allow for higher throughput with multiple parallel requests but the results will **not** be deterministic due to differences in rounding error. -- `-cb`, `--cont-batching`: Enable continuous batching (a.k.a dynamic batching). Default: disabled -- `-spf FNAME`, `--system-prompt-file FNAME` Set a file to load a system prompt (initial prompt of all slots). This is useful for chat applications. [See more](#change-system-prompt-on-runtime) -- `--mmproj MMPROJ_FILE`: Path to a multimodal projector file for LLaVA. -- `--grp-attn-n`: Set the group attention factor to extend context size through self-extend. Used together with group attention width `--grp-attn-w`. Default: `1`, which is disabled. -- `--grp-attn-w`: Set the group attention width to extend context size through self-extend. Used together with group attention factor `--grp-attn-n`. Default: `512` -- `-n N, --n-predict N`: Set the maximum tokens to predict. Default: `-1` -- `--slots-endpoint-disable`: To disable slots state monitoring endpoint. Slots state may contain user data, prompts included. -- `--metrics`: enable prometheus `/metrics` compatible endpoint. Default: disabled -- `--slot-save-path PATH`: Specifies the path where the state of slots (the prompt cache) can be stored. If not provided, the slot management endpoints will be disabled. -- `--chat-template JINJA_TEMPLATE`: Set custom jinja chat template. This parameter accepts a string, not a file name. Default: template taken from model's metadata. We only support [some pre-defined templates](https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template) -- `--log-disable`: Output logs to stdout only, not to `llama.log`. Default: enabled -- `--log-format FORMAT`: Define the log output to FORMAT: json or text Default: `json` -- `--rope-scaling` : RoPE scaling method. Defaults to linear unless otherwise specified by the model. Options are `none`, `linear`, `yarn` -- `--rope-freq-base N` : RoPE frequency base (default: loaded from model) -- `--rope-freq-scale N`: RoPE frequency scaling factor, expands context by a factor of 1/N (e.g. 0.25) -- `--yarn-ext-factor N` : YaRN: extrapolation mix factor (Default: 1.0, 0.0 = full interpolation) -- `--yarn-attn-factor N` : YaRN: scale sqrt(t) or attention magnitude (default: 1.0) -- `--yarn-beta-slow N`: YaRN: High correction dim or alpha (default: 1.0) -- `--yarn-beta-fast N`: YaRN: low correction dim or beta (default: 32.0) -- `--pooling` : Pooling type for embeddings, use model default if unspecified. Options are `none`, `mean`, `cls` -- `-dt N`, `--defrag-thold N`: KV cache defragmentation threshold (default: -1.0, < 0 = disabled) -- `-fa`, `--flash-attn` : enable flash attention (default: disabled). -- `-ctk TYPE`, `--cache-type-k TYPE` : KV cache data type for K (default: `f16`, options `f32`, `f16`, `q8_0`, `q4_0`, `q4_1`, `iq4_nl`, `q5_0`, or `q5_1`) -- `-ctv TYPE`, `--cache-type-v TYPE` : KV cache type for V (default `f16`, see `-ctk` for options) +``` +usage: ./llama-server [options] + +general: + + -h, --help, --usage print usage and exit + --version show version and build info + -v, --verbose print verbose information + --verbosity N set specific verbosity level (default: 0) + --verbose-prompt print a verbose prompt before generation (default: false) + --no-display-prompt don't print prompt at generation (default: false) + -co, --color colorise output to distinguish prompt and user input from generations (default: false) + -s, --seed SEED RNG seed (default: -1, use random seed for < 0) + -t, --threads N number of threads to use during generation (default: 8) + -tb, --threads-batch N number of threads to use during batch and prompt processing (default: same as --threads) + -td, --threads-draft N number of threads to use during generation (default: same as --threads) + -tbd, --threads-batch-draft N number of threads to use during batch and prompt processing (default: same as --threads-draft) + --draft N number of tokens to draft for speculative decoding (default: 5) + -ps, --p-split N speculative decoding split probability (default: 0.1) + -lcs, --lookup-cache-static FNAME + path to static lookup cache to use for lookup decoding (not updated by generation) + -lcd, --lookup-cache-dynamic FNAME + path to dynamic lookup cache to use for lookup decoding (updated by generation) + -c, --ctx-size N size of the prompt context (default: 0, 0 = loaded from model) + -n, --predict N number of tokens to predict (default: -1, -1 = infinity, -2 = until context filled) + -b, --batch-size N logical maximum batch size (default: 2048) + -ub, --ubatch-size N physical maximum batch size (default: 512) + --keep N number of tokens to keep from the initial prompt (default: 0, -1 = all) + --chunks N max number of chunks to process (default: -1, -1 = all) + -fa, --flash-attn enable Flash Attention (default: disabled) + -p, --prompt PROMPT prompt to start generation with + in conversation mode, this will be used as system prompt + (default: '') + -f, --file FNAME a file containing the prompt (default: none) + --in-file FNAME an input file (repeat to specify multiple files) + -bf, --binary-file FNAME binary file containing the prompt (default: none) + -e, --escape process escapes sequences (\n, \r, \t, \', \", \\) (default: true) + --no-escape do not process escape sequences + -ptc, --print-token-count N print token count every N tokens (default: -1) + --prompt-cache FNAME file to cache prompt state for faster startup (default: none) + --prompt-cache-all if specified, saves user input and generations to cache as well + not supported with --interactive or other interactive options + --prompt-cache-ro if specified, uses the prompt cache but does not update it + -r, --reverse-prompt PROMPT halt generation at PROMPT, return control in interactive mode + can be specified more than once for multiple prompts + -sp, --special special tokens output enabled (default: false) + -cnv, --conversation run in conversation mode, does not print special tokens and suffix/prefix + if suffix/prefix are not specified, default chat template will be used + (default: false) + -i, --interactive run in interactive mode (default: false) + -if, --interactive-first run in interactive mode and wait for input right away (default: false) + -mli, --multiline-input allows you to write or paste multiple lines without ending each in '\' + --in-prefix-bos prefix BOS to user inputs, preceding the `--in-prefix` string + --in-prefix STRING string to prefix user inputs with (default: empty) + --in-suffix STRING string to suffix after user inputs with (default: empty) + --spm-infill use Suffix/Prefix/Middle pattern for infill (instead of Prefix/Suffix/Middle) as some models prefer this. (default: disabled) + +sampling: + + --samplers SAMPLERS samplers that will be used for generation in the order, separated by ';' + (default: top_k;tfs_z;typical_p;top_p;min_p;temperature) + --sampling-seq SEQUENCE simplified sequence for samplers that will be used (default: kfypmt) + --ignore-eos ignore end of stream token and continue generating (implies --logit-bias EOS-inf) + --penalize-nl penalize newline tokens (default: false) + --temp N temperature (default: 0.8) + --top-k N top-k sampling (default: 40, 0 = disabled) + --top-p N top-p sampling (default: 0.9, 1.0 = disabled) + --min-p N min-p sampling (default: 0.1, 0.0 = disabled) + --tfs N tail free sampling, parameter z (default: 1.0, 1.0 = disabled) + --typical N locally typical sampling, parameter p (default: 1.0, 1.0 = disabled) + --repeat-last-n N last n tokens to consider for penalize (default: 64, 0 = disabled, -1 = ctx_size) + --repeat-penalty N penalize repeat sequence of tokens (default: 1.0, 1.0 = disabled) + --presence-penalty N repeat alpha presence penalty (default: 0.0, 0.0 = disabled) + --frequency-penalty N repeat alpha frequency penalty (default: 0.0, 0.0 = disabled) + --dynatemp-range N dynamic temperature range (default: 0.0, 0.0 = disabled) + --dynatemp-exp N dynamic temperature exponent (default: 1.0) + --mirostat N use Mirostat sampling. + Top K, Nucleus, Tail Free and Locally Typical samplers are ignored if used. + (default: 0, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0) + --mirostat-lr N Mirostat learning rate, parameter eta (default: 0.1) + --mirostat-ent N Mirostat target entropy, parameter tau (default: 5.0) + -l TOKEN_ID(+/-)BIAS modifies the likelihood of token appearing in the completion, + i.e. `--logit-bias 15043+1` to increase likelihood of token ' Hello', + or `--logit-bias 15043-1` to decrease likelihood of token ' Hello' + --cfg-negative-prompt PROMPT + negative prompt to use for guidance (default: '') + --cfg-negative-prompt-file FNAME + negative prompt file to use for guidance + --cfg-scale N strength of guidance (default: 1.0, 1.0 = disable) + --chat-template JINJA_TEMPLATE + set custom jinja chat template (default: template taken from model's metadata) + if suffix/prefix are specified, template will be disabled + only commonly used templates are accepted: + https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template + +grammar: + + --grammar GRAMMAR BNF-like grammar to constrain generations (see samples in grammars/ dir) (default: '') + --grammar-file FNAME file to read grammar from + -j, --json-schema SCHEMA JSON schema to constrain generations (https://json-schema.org/), e.g. `{}` for any JSON object + For schemas w/ external $refs, use --grammar + example/json_schema_to_grammar.py instead + +embedding: + + --pooling {none,mean,cls,last} + pooling type for embeddings, use model default if unspecified + --attention {causal,non-causal} + attention type for embeddings, use model default if unspecified + +context hacking: + + --rope-scaling {none,linear,yarn} + RoPE frequency scaling method, defaults to linear unless specified by the model + --rope-scale N RoPE context scaling factor, expands context by a factor of N + --rope-freq-base N RoPE base frequency, used by NTK-aware scaling (default: loaded from model) + --rope-freq-scale N RoPE frequency scaling factor, expands context by a factor of 1/N + --yarn-orig-ctx N YaRN: original context size of model (default: 0 = model training context size) + --yarn-ext-factor N YaRN: extrapolation mix factor (default: -1.0, 0.0 = full interpolation) + --yarn-attn-factor N YaRN: scale sqrt(t) or attention magnitude (default: 1.0) + --yarn-beta-slow N YaRN: high correction dim or alpha (default: 1.0) + --yarn-beta-fast N YaRN: low correction dim or beta (default: 32.0) + -gan, --grp-attn-n N group-attention factor (default: 1) + -gaw, --grp-attn-w N group-attention width (default: 512.0) + -dkvc, --dump-kv-cache verbose print of the KV cache + -nkvo, --no-kv-offload disable KV offload + -ctk, --cache-type-k TYPE KV cache data type for K (default: f16) + -ctv, --cache-type-v TYPE KV cache data type for V (default: f16) + +perplexity: + + --all-logits return logits for all tokens in the batch (default: false) + --hellaswag compute HellaSwag score over random tasks from datafile supplied with -f + --hellaswag-tasks N number of tasks to use when computing the HellaSwag score (default: 400) + --winogrande compute Winogrande score over random tasks from datafile supplied with -f + --winogrande-tasks N number of tasks to use when computing the Winogrande score (default: 0) + --multiple-choice compute multiple choice score over random tasks from datafile supplied with -f + --multiple-choice-tasks N + number of tasks to use when computing the multiple choice score (default: 0) + --kl-divergence computes KL-divergence to logits provided via --kl-divergence-base + --ppl-stride N stride for perplexity calculation (default: 0) + --ppl-output-type {0,1} output type for perplexity calculation (default: 0) + +parallel: + + -dt, --defrag-thold N KV cache defragmentation threshold (default: -1.0, < 0 - disabled) + -np, --parallel N number of parallel sequences to decode (default: 1) + -ns, --sequences N number of sequences to decode (default: 1) + -cb, --cont-batching enable continuous batching (a.k.a dynamic batching) (default: enabled) + +multi-modality: + + --mmproj FILE path to a multimodal projector file for LLaVA. see examples/llava/README.md + --image FILE path to an image file. use with multimodal models. Specify multiple times for batching + +backend: + + --rpc SERVERS comma separated list of RPC servers + --mlock force system to keep model in RAM rather than swapping or compressing + --no-mmap do not memory-map model (slower load but may reduce pageouts if not using mlock) + --numa TYPE attempt optimizations that help on some NUMA systems + - distribute: spread execution evenly over all nodes + - isolate: only spawn threads on CPUs on the node that execution started on + - numactl: use the CPU map provided by numactl + if run without this previously, it is recommended to drop the system page cache before using this + see https://github.com/ggerganov/llama.cpp/issues/1437 + +model: + + --check-tensors check model tensor data for invalid values (default: false) + --override-kv KEY=TYPE:VALUE + advanced option to override model metadata by key. may be specified multiple times. + types: int, float, bool, str. example: --override-kv tokenizer.ggml.add_bos_token=bool:false + --lora FNAME apply LoRA adapter (implies --no-mmap) + --lora-scaled FNAME S apply LoRA adapter with user defined scaling S (implies --no-mmap) + --lora-base FNAME optional model to use as a base for the layers modified by the LoRA adapter + --control-vector FNAME add a control vector + note: this argument can be repeated to add multiple control vectors + --control-vector-scaled FNAME SCALE + add a control vector with user defined scaling SCALE + note: this argument can be repeated to add multiple scaled control vectors + --control-vector-layer-range START END + layer range to apply the control vector(s) to, start and end inclusive + -m, --model FNAME model path (default: models/$filename with filename from --hf-file + or --model-url if set, otherwise models/7B/ggml-model-f16.gguf) + -md, --model-draft FNAME draft model for speculative decoding (default: unused) + -mu, --model-url MODEL_URL model download url (default: unused) + -hfr, --hf-repo REPO Hugging Face model repository (default: unused) + -hff, --hf-file FILE Hugging Face model file (default: unused) + -hft, --hf-token TOKEN Hugging Face access token (default: value from HF_TOKEN environment variable) + +retrieval: + + --context-file FNAME file to load context from (repeat to specify multiple files) + --chunk-size N minimum length of embedded text chunks (default: 64) + --chunk-separator STRING + separator between chunks (default: ' + ') + +passkey: + + --junk N number of times to repeat the junk text (default: 250) + --pos N position of the passkey in the junk text (default: -1) + +imatrix: + + -o, --output FNAME output file (default: 'imatrix.dat') + --output-frequency N output the imatrix every N iterations (default: 10) + --save-frequency N save an imatrix copy every N iterations (default: 0) + --process-output collect data for the output tensor (default: false) + --no-ppl do not compute perplexity (default: true) + --chunk N start processing the input from chunk N (default: 0) + +bench: + + -pps is the prompt shared across parallel sequences (default: false) + -npp n0,n1,... number of prompt tokens + -ntg n0,n1,... number of text generation tokens + -npl n0,n1,... number of parallel prompts + +embedding: + + --embd-normalize normalisation for embendings (default: 2) (-1=none, 0=max absolute int16, 1=taxicab, 2=euclidean, >2=p-norm) + --embd-output-format empty = default, "array" = [[],[]...], "json" = openai style, "json+" = same "json" + cosine similarity matrix + --embd-separator separator of embendings (default \n) for example "<#sep#>" + +server: + + --host HOST ip address to listen (default: 127.0.0.1) + --port PORT port to listen (default: 8080) + --path PATH path to serve static files from (default: ) + --embedding(s) enable embedding endpoint (default: disabled) + --api-key KEY API key to use for authentication (default: none) + --api-key-file FNAME path to file containing API keys (default: none) + --ssl-key-file FNAME path to file a PEM-encoded SSL private key + --ssl-cert-file FNAME path to file a PEM-encoded SSL certificate + --timeout N server read/write timeout in seconds (default: 600) + --threads-http N number of threads used to process HTTP requests (default: -1) + --system-prompt-file FNAME + set a file to load a system prompt (initial prompt of all slots), this is useful for chat applications + --log-format {text,json} + log output format: json or text (default: json) + --metrics enable prometheus compatible metrics endpoint (default: disabled) + --no-slots disables slots monitoring endpoint (default: enabled) + --slot-save-path PATH path to save slot kv cache (default: disabled) + --chat-template JINJA_TEMPLATE + set custom jinja chat template (default: template taken from model's metadata) + only commonly used templates are accepted: + https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template + -sps, --slot-prompt-similarity SIMILARITY + how much the prompt of a request must match the prompt of a slot in order to use that slot (default: 0.50, 0.0 = disabled) + + +logging: + + --simple-io use basic IO for better compatibility in subprocesses and limited consoles + -ld, --logdir LOGDIR path under which to save YAML logs (no logging if unset) + --log-test Run simple logging test + --log-disable Disable trace logs + --log-enable Enable trace logs + --log-file FNAME Specify a log filename (without extension) + --log-new Create a separate new log file on start. Each log file will have unique name: "..log" + --log-append Don't truncate the old log file. + +cvector: + + -o, --output FNAME output file (default: 'control_vector.gguf') + --positive-file FNAME positive prompts file, one prompt per line (default: 'examples/cvector-generator/positive.txt') + --negative-file FNAME negative prompts file, one prompt per line (default: 'examples/cvector-generator/negative.txt') + --pca-batch N batch size used for PCA. Larger batch runs faster, but uses more memory (default: 100) + --pca-iter N number of iterations used for PCA (default: 1000) + --method {pca,mean} dimensionality reduction method to be used (default: pca) +``` -**If compiled with `LLAMA_SERVER_SSL=ON`** -- `--ssl-key-file FNAME`: path to file a PEM-encoded SSL private key -- `--ssl-cert-file FNAME`: path to file a PEM-encoded SSL certificate ## Build @@ -231,7 +444,7 @@ node index.js `n_predict`: Set the maximum number of tokens to predict when generating text. **Note:** May exceed the set limit slightly if the last token is a partial multibyte character. When 0, no tokens will be generated but the prompt is evaluated into the cache. Default: `-1`, where `-1` is infinity. - `n_keep`: Specify the number of tokens from the prompt to retain when the context size is exceeded and tokens need to be discarded. + `n_keep`: Specify the number of tokens from the prompt to retain when the context size is exceeded and tokens need to be discarded. The number excludes the BOS token. By default, this value is set to `0`, meaning no tokens are kept. Use `-1` to retain all tokens from the prompt. `stream`: It allows receiving each predicted token in real-time instead of waiting for the completion to finish. To enable this, set to `true`. @@ -365,7 +578,8 @@ Notice that each `probs` is an array of length `n_probs`. "assistant_name": "", "user_name": "", "default_generation_settings": { ... }, - "total_slots": 1 + "total_slots": 1, + "chat_template": "" } ``` @@ -373,8 +587,9 @@ Notice that each `probs` is an array of length `n_probs`. - `user_name` - the required anti-prompt to generate the prompt in case you have specified a system prompt for all slots. - `default_generation_settings` - the default generation settings for the `/completion` endpoint, which has the same fields as the `generation_settings` response object from the `/completion` endpoint. - `total_slots` - the total number of slots for process requests (defined by `--parallel` option) +- `chat_template` - the model's original Jinja2 prompt template -- **POST** `/v1/chat/completions`: OpenAI-compatible Chat Completions API. Given a ChatML-formatted json description in `messages`, it returns the predicted completion. Both synchronous and streaming mode are supported, so scripted and interactive applications work fine. While no strong claims of compatibility with OpenAI API spec is being made, in our experience it suffices to support many apps. Only model with [supported chat template](https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template) can be used optimally with this endpoint. By default, ChatML template will be used. +- **POST** `/v1/chat/completions`: OpenAI-compatible Chat Completions API. Given a ChatML-formatted json description in `messages`, it returns the predicted completion. Both synchronous and streaming mode are supported, so scripted and interactive applications work fine. While no strong claims of compatibility with OpenAI API spec is being made, in our experience it suffices to support many apps. Only models with a [supported chat template](https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template) can be used optimally with this endpoint. By default, the ChatML template will be used. *Options:* diff --git a/examples/server/bench/bench.py b/examples/server/bench/bench.py index 4fbbb2032..2daac0884 100644 --- a/examples/server/bench/bench.py +++ b/examples/server/bench/bench.py @@ -1,3 +1,5 @@ +from __future__ import annotations + import argparse import json import os @@ -59,10 +61,11 @@ def main(args_in: list[str] | None = None) -> None: sys.exit(1) # start the benchmark + iterations = 0 + data = {} try: start_benchmark(args) - iterations = 0 with open("results.github.env", 'w') as github_env: # parse output with open('k6-results.json', 'r') as bench_results: @@ -129,7 +132,7 @@ def main(args_in: list[str] | None = None) -> None: timestamps, metric_values = zip(*values) metric_values = [float(value) for value in metric_values] prometheus_metrics[metric] = metric_values - timestamps_dt = [datetime.fromtimestamp(int(ts)) for ts in timestamps] + timestamps_dt = [str(datetime.fromtimestamp(int(ts))) for ts in timestamps] plt.figure(figsize=(16, 10), dpi=80) plt.plot(timestamps_dt, metric_values, label=metric) plt.xticks(rotation=0, fontsize=14, horizontalalignment='center', alpha=.7) @@ -156,7 +159,7 @@ def main(args_in: list[str] | None = None) -> None: plt.close() # Mermaid format in case images upload failed - with (open(f"{metric}.mermaid", 'w') as mermaid_f): + with open(f"{metric}.mermaid", 'w') as mermaid_f: mermaid = ( f"""--- config: @@ -278,7 +281,7 @@ def start_server_background(args): } server_process = subprocess.Popen( args, - **pkwargs) + **pkwargs) # pyright: ignore[reportArgumentType, reportCallIssue] def server_log(in_stream, out_stream): for line in iter(in_stream.readline, b''): diff --git a/examples/server/public/completion.js b/examples/server/public/completion.js index 987b9a3b8..36818f764 100644 --- a/examples/server/public/completion.js +++ b/examples/server/public/completion.js @@ -21,7 +21,7 @@ let generation_settings = null; // export async function* llama(prompt, params = {}, config = {}) { let controller = config.controller; - const api_url = config.api_url || ""; + const api_url = config.api_url?.replace(/\/+$/, '') || ""; if (!controller) { controller = new AbortController(); @@ -196,7 +196,7 @@ export const llamaComplete = async (params, controller, callback) => { // Get the model info from the server. This is useful for getting the context window and so on. export const llamaModelInfo = async (config = {}) => { if (!generation_settings) { - const api_url = config.api_url || ""; + const api_url = config.api_url?.replace(/\/+$/, '') || ""; const props = await fetch(`${api_url}/props`).then(r => r.json()); generation_settings = props.default_generation_settings; } diff --git a/examples/server/public/index-new.html b/examples/server/public/index-new.html index 5513e9121..c87dd8f1e 100644 --- a/examples/server/public/index-new.html +++ b/examples/server/public/index-new.html @@ -14,10 +14,10 @@