mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-27 03:44:35 +00:00

History

Xuan Son Nguyen cda0e4b648 Some checks failed Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/full-cuda.Dockerfile platforms:linux/amd64 tag:full-cuda]) (push) Has been cancelled Details Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/full-musa.Dockerfile platforms:linux/amd64 tag:full-musa]) (push) Has been cancelled Details Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/full.Dockerfile platforms:linux/amd64,linux/arm64 tag:full]) (push) Has been cancelled Details Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/llama-cli-cuda.Dockerfile platforms:linux/amd64 tag:light-cuda]) (push) Has been cancelled Details Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/llama-cli-intel.Dockerfile platforms:linux/amd64 tag:light-intel]) (push) Has been cancelled Details Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/llama-cli-musa.Dockerfile platforms:linux/amd64 tag:light-musa]) (push) Has been cancelled Details Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/llama-cli.Dockerfile platforms:linux/amd64,linux/arm64 tag:light]) (push) Has been cancelled Details Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/llama-server-cuda.Dockerfile platforms:linux/amd64 tag:server-cuda]) (push) Has been cancelled Details Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/llama-server-intel.Dockerfile platforms:linux/amd64 tag:server-intel]) (push) Has been cancelled Details Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/llama-server-musa.Dockerfile platforms:linux/amd64 tag:server-musa]) (push) Has been cancelled Details Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/llama-server.Dockerfile platforms:linux/amd64,linux/arm64 tag:server]) (push) Has been cancelled Details Nix CI / nix-eval (macos-latest) (push) Has been cancelled Details Nix CI / nix-eval (ubuntu-latest) (push) Has been cancelled Details Nix CI / nix-build (macos-latest) (push) Has been cancelled Details Nix CI / nix-build (ubuntu-latest) (push) Has been cancelled Details flake8 Lint / Lint (push) Has been cancelled Details update-flake-lock / lockfile (push) Has been cancelled Details llama : remove all_pos_0, all_pos_1, all_seq_id from llama_batch (#9745 ) * refactor llama_batch_get_one * adapt all examples * fix simple.cpp * fix llama_bench * fix * fix context shifting * free batch before return * use common_batch_add, reuse llama_batch in loop * null terminated seq_id list * fix save-load-state example * fix perplexity * correct token pos in llama_batch_allocr		2024-10-18 23:18:01 +02:00
..
CMakeLists.txt	`build`: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (#7809 )	2024-06-13 00:41:52 +01:00
imatrix.cpp	llama : remove all_pos_0, all_pos_1, all_seq_id from llama_batch (#9745 )	2024-10-18 23:18:01 +02:00
README.md	docs : Quantum -> Quantized (#8666 )	2024-07-25 11:13:27 +03:00

README.md

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 quantized models. More information is available here: https://github.com/ggerganov/llama.cpp/pull/4861

Usage

./llama-imatrix \
    -m model.gguf -f some-text.txt [-o imatrix.dat] [--process-output] [--verbosity 1] \
    [--no-ppl] [--chunk 123] [--output-frequency 10] [--save-frequency 0] \
    [--in-file imatrix-prev-0.dat --in-file imatrix-prev-1.dat ...]

Here -m with a model name and -f with a file containing training data (such as e.g. wiki.train.raw) are mandatory. The parameters in square brackets are optional and have the following meaning:

-o (or --output-file) specifies the name of the file where the computed data will be stored. If missing imatrix.dat is used.
--verbosity specifies the verbosity level. If set to 0, no output other than the perplexity of the processed chunks will be generated. If set to 1, each time the results are saved a message is written to stderr. If >=2, a message is output each time data is collected for any tensor. Default verbosity level is 1.
--output-frequency specifies how often the so far computed result is saved to disk. Default is 10 (i.e., every 10 chunks)
--save-frequency specifies how often to save a copy of the imatrix in a separate file. Default is 0 (i.e., never)
--process-output specifies if data will be collected for the output.weight tensor. My experience is that it is better to not utilize the importance matrix when quantizing output.weight, so this is set to false by default.

For faster computation, make sure to use GPU offloading via the -ngl argument

Example

GGML_CUDA=1 make -j

# generate importance matrix (imatrix.dat)
./llama-imatrix -m ggml-model-f16.gguf -f train-data.txt -ngl 99

# use the imatrix to perform a Q4_K_M quantization
./llama-quantize --imatrix imatrix.dat ggml-model-f16.gguf ./ggml-model-q4_k_m.gguf q4_k_m