Commit Graph

198 Commits

Author SHA1 Message Date
Johannes Gäßler
a743d76a01
CUDA: generalize FP16 fattn vec kernel (#7061)
* CUDA: generalize FP16 fattn vec kernel

* disable unsupported head sizes for AMD in test

* try AMD fix

* fix batch size 2-8

* partially revert changes
2024-05-09 14:32:02 +02:00
Johannes Gäßler
c12452c7ae
JSON: [key] -> .at(key), assert() -> GGML_ASSERT (#7143) 2024-05-08 21:53:08 +02:00
Ren Xuancheng
229ffff872
llama : add BPE pre-tokenization for Qwen2 (#7114)
* Add BPE pre-tokenization for Qwen2.

* minor : fixes

---------

Co-authored-by: Ren Xuancheng <17811943+jklj077@users.noreply.github.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-05-08 15:06:43 +03:00
Justine Tunney
3855416027
ggml : introduce bfloat16 support (#6412)
* Introduce bfloat16 support

Many models on Hugging Face (e.g. Mistral, TinyLLaMA) use bfloat16 as
their canonical floating point format.

      ┌sign
      │
      │   ┌exponent
      │   │
      │   │      ┌mantissa
      │   │      │
      │┌──┴───┐┌─┴───┐
    0b0000000000000000 brain16

This encoding has the same number of exponent bits as float32. That
makes conversion relatively straightforward, even in the absence of
hardware support. For example, converting brain16 to binary32 means
simply shifting 16 bits to the left.

      ┌sign
      │
      │   ┌exponent
      │   │
      │   │      ┌mantissa
      │   │      │
      │┌──┴───┐┌─┴───────────────────┐
    0b00000000000000000000000000000000 IEEE binary32

The issue is that converting bf16 to fp16 can result in information
loss. Only 13% of bf16 numbers can be precisely represented in fp16
which in practice ends up being 99.71% of Mistral 7b v0.2's weights
however there is currently no way other than fp32 to get the others

      ┌sign
      │
      │  ┌exponent
      │  │
      │  │    ┌mantissa
      │  │    │
      │┌─┴─┐┌─┴──────┐
    0b0000000000000000 IEEE binary16

This change fixes that, by adding a bf16 data type to GGML. Support
for CPU inference has been implemented along with optimizations for
the AVX2, AVX512, and AVX512BF16 ISAs. Perplexity on Mistral 7b 0.2
improves somewhere around -0.0024 to -0.0046 compared to using fp16

* Remove GGML code that's not needed

* Minimize the GGML API surface area for BF16

* Remove bf16 luts

* Make the GGML header look nicer

* Fix documentation

* Apply ggerganov's fixes for test-backend-ops

* Add BF16 code for new ggml_validate_row_data() function
2024-05-08 09:30:09 +03:00
DAN™
889bdd7686
command-r : add BPE pre-tokenization (#7063)
* Add BPE pre-tokenization for Command-R/R+.

* Bump transformers convert requirement.

* command-r : add individual digits regex

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-05-05 08:19:30 +03:00
Brian
6fbd432211
py : logging and flake8 suppression refactoring (#7081)
Set one as executable and add basicConfig()
to another. Also added noqa tag to test scripts.
2024-05-05 08:07:48 +03:00
Georgi Gerganov
92139b90af
tests : add test-tokenizer-0.sh + fix some tokenizers (#7036)
* tests : add test-tokenizer-0.sh

* unicode : add all unicode number ranges

* starcoder : fix pre-tokenizer

* tests : add test that fails with DeepSeek tokenizers

* falcon : fix regex

* unicode : regenerate unicode tables

* refact : add tokenizer model

* lint : fix

* tests : disable failing tests

ggml-ci

* refact : add tests files

ggml-ci

* convert : print -> logging

ggml-ci

* lint : fix

* unicode : digit -> number

* phi-3 : update
2024-05-04 08:32:32 +03:00
Brian
a2ac89d6ef
convert.py : add python logging instead of print() (#6511)
* convert.py: add python logging instead of print()

* convert.py: verbose flag takes priority over dump flag log suppression

* convert.py: named instance logging

* convert.py: use explicit logger id string

* convert.py: convert extra print() to named logger

* convert.py: sys.stderr.write --> logger.error

* *.py: Convert all python scripts to use logging module

* requirements.txt: remove extra line

* flake8: update flake8 ignore and exclude to match ci settings

* gh-actions: add flake8-no-print to flake8 lint step

* pre-commit: add flake8-no-print to flake8 and also update pre-commit version

* convert-hf-to-gguf.py: print() to logger conversion

* *.py: logging basiconfig refactor to use conditional expression

* *.py: removed commented out logging

* fixup! *.py: logging basiconfig refactor to use conditional expression

* constant.py: logger.error then exit should be a raise exception instead

* *.py: Convert logger error and sys.exit() into a raise exception (for atypical error)

* gguf-convert-endian.py: refactor convert_byteorder() to use tqdm progressbar

* verify-checksum-model.py: This is the result of the program, it should be printed to stdout.

* compare-llama-bench.py: add blank line for readability during missing repo response

* reader.py: read_gguf_file() use print() over logging

* convert.py: warning goes to stderr and won't hurt the dump output

* gguf-dump.py: dump_metadata() should print to stdout

* convert-hf-to-gguf.py: print --> logger.debug or ValueError()

* verify-checksum-models.py: use print() for printing table

* *.py: refactor logging.basicConfig()

* gguf-py/gguf/*.py: use __name__ as logger name

Since they will be imported and not run directly.

* python-lint.yml: use .flake8 file instead

* constants.py: logger no longer required

* convert-hf-to-gguf.py: add additional logging

* convert-hf-to-gguf.py: print() --> logger

* *.py: fix flake8 warnings

* revert changes to convert-hf-to-gguf.py for get_name()

* convert-hf-to-gguf-update.py: use triple quoted f-string instead

* *.py: accidentally corrected the wrong line

* *.py: add compilade warning suggestions and style fixes
2024-05-03 22:36:41 +03:00
Georgi Gerganov
9c67c2773d
ggml : add Flash Attention (#5021)
* ggml : add ggml_flash_attn_ext API

* ggml : fix GQA support in ggml_flash_attn_ext

* ggml : online attention (CPU)

* metal : initial implementation

* metal : f16 precision

* metal : reduce branches

* metal : specialize for head size

* wip : 8 rows per simd group

* wip : 4 rows per simd group

* wip : template for rows per warp

* metal : parallelize across KV size

* metal : parallel reduce across heads

* metal : efficient flash_attn_f16 implementation

* metal : avoid redundant loads of the attention

* metal : scale and mask in matrix form

* metal : fix comment

* llama : avoid ggml_cast, use F32 query

* metal : add parallel reduce version (disabled)

* metal : move output into local memory + optimize

- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments

* metal : add tests, fix scaling, support C > 32

* metal : improve precision

* ggml : fix f16 mad

* metal : minor

* metal : support Q > 8

* tests : add ATTN tests

* metal : disable buffer allocation logs

* tests : more

* metal : faster inner loop for C == 32

* metal : fix array initialization

* tests : ifdef

* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext

* ggml : fix ggml_soft_max mask requirement

* cuda : fix soft_max to use correct mask size

* cuda : add flash_attn kernel (wip)

* metal : optimize softmax for C > 32

* metal : optimize softmax

* tests : minor fix

* cuda : avoid zeroing fragments

* tests : update dims

* cuda : fix __hisinf() result check

* cuda : avoid warp_reduce for smax

* cuda : use int instead of int64_t

Noticeably improves performance (thanks to Johannes)

* cuda : make loops use the same loop values

Thanks Johannes again for the tip

* cuda : unroll some of the loops

* cuda : avoid __hisinf branches

* cuda : use half2 in softmax

* cuda : switch to 1 warp for bs > 16

* cuda : speed-up reduce part of the kernel

* cuda : unroll Q*K^T loop

* cuda : fix -INF block check

* cuda : simplify softmax

* cuda : fix matrix names

* cuda : minor

* llama : adapt to F16 KQ_pos

* llama : adapt new models to F16 KQ_mask

* ggml : fix F16 store (ARM NEON)

* llama : fix type of KQ_mask and KQ_pos

* ggml : fix CPU soft_max

* tests : add hs=256

* cuda : fix build

* metal : improve perf via smaller int registers

* cuda : adapt soft_max to F16 mask and pos

* CUDA: faster FlashAttention, kernel for bs == 1

* 16 cols for Phi-2

* no vec for hs, no hs==256 ncols==32 for Volta

* adjust kernel selection logic

* 4 warps, 256 stride for all D

* no ncols == 64

* Multiple parallel blocks for batch size 1

* fix compile warnings

* fix excessive KQ_b loads

* fix cmake build

* fix KV cache padding, NaN from INFINITY (#6438)

* llama : flash_attn cparam + fix defrag

* server: support flash_attn param

* server: bench: enable flash_attn param

* CUDA: refactor host code, dyn. par. blocks

* fix flash_attn_vec_f16 race condition

* flush softmax exp below threshold to 0

* store temp KQ in registers

* Calculate KQ as FP32 if KQV has GGML_PREC_F32

* Add __hgt2_mask implementation for CUDA 11

* fix KQ FP32 precision fpr parallel_blocks > 1

* llama-bench : add -fa,--flash-attn arg

* metal : add BS=1 kernel for flash attention (#6508)

* metal : add BS=1 kernel for flash attention (wip)

* metal : support more than 1 warps

* metal : opts

* metal : opt

* metal : switch to parallel reduce

* metal : reduce registers

* metal : simplify

* metal : initial FA vec kernel

* metal : use F32 attention accumulators

* batched-bench : add fattn arg

* llama : simplify llama_build_kv_store

ggml-ci

* llama : adapt build_olmo to changes

* ggml : fix arm fp16 store on windows

* metal : clean-up

* metal : clean-up kernel code

* metal : minor

* tests : remove benchmarks

ggml-ci

* ggml : fix avx512 const correctness

ggml-ci

* ggml : fix soft_max with bias on CPU

ggml-ci

* common : print --flash-attn in help

* ggml : fix num dimensions in ggml_flash_attn_ext

* llama : force disable flash attention for incompatible models

* ggml : ggml_soft_max support F16/F32 mask/pos

ggml-ci

* cuda : uint -> uint32_t

* cuda : "constexpr dim3" -> "const dim3"

ggml-ci

* cuda : try to fix __hgt2_mask

ggml-ci

* ggml : add TODO's for F16/F32 mask/pos support in other backends

* llama : replace bool need_kq_pos with use_alibi

* llama : prep ALiBi support for BERT models

ggml-ci

* llama : fix n_batch requirements

ggml-ci

* cont

* server : add help for --flash-attn arg

* llama : disable FA for AMD

* tests : remove TMP_ATTN_BENCH

ggml-ci

* llama : support save/load state with FA enabled

ggml-ci

* ci : add CUDA save-load-state tests

ggml-ci

* llama : llama_kv_cache_clear zeroes data + fix save-load seq

ggml-ci

* llama : fix copy-paste errors, add TODO

* llama : disallow incompatible states

* llama : update llama_state_get_size after v_trans field

* metal : remove tmp log

* llama : add static reminder for llama_state_get_size

* metal : fix max nsg

ggml-ci

* ci : fix arg order

ggml-ci

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 12:16:08 +03:00
Clint Herron
b8c1476e44
Extending grammar integration tests (#6644)
* Cleaning up integration tests to share code between tests and make it simpler to add new tests.

* Add tests around quantifiers to ensure both matching and non-matching compliance.

* Add slightly more complex grammar with quantifiers to test references with quantifiers.

* Fixing build when C++17 is not present.

* Separating test calls to give more helpful stack traces on failure. Adding verbose messages to give visibility for what is being tested.

* Adding quotes around strings to explicitly show whitespace

* Removing trailing whitespace.

* Implementing suggestions from @ochafik -- grammars and test strings now print and flush before tests to aid in debugging segfaults and whatnot.

* Cleaning up forgotten symbols. Modifying simple test to use test harness. Added comments for more verbose descriptions of what each test is accomplishing.

* Unicode symbol modifications to hopefully make log easier to parse visually.
2024-04-29 14:40:14 -04:00
Georgi Gerganov
f4ab2a4147
llama : fix BPE pre-tokenization (#6920)
* merged the changes from deepseeker models to main branch

* Moved regex patterns to unicode.cpp and updated unicode.h

* Moved header files

* Resolved issues

* added and refactored unicode_regex_split and related functions

* Updated/merged the deepseek coder pr

* Refactored code

* Adding unicode regex mappings

* Adding unicode regex function

* Added needed functionality, testing remains

* Fixed issues

* Fixed issue with gpt2 regex custom preprocessor

* unicode : fix? unicode_wstring_to_utf8

* lint : fix whitespaces

* tests : add tokenizer tests for numbers

* unicode : remove redundant headers

* tests : remove and rename tokenizer test scripts

* tests : add sample usage

* gguf-py : reader prints warnings on duplicate keys

* llama : towards llama3 tokenization support (wip)

* unicode : shot in the dark to fix tests on Windows

* unicode : first try custom implementations

* convert : add "tokenizer.ggml.pre" GGUF KV (wip)

* llama : use new pre-tokenizer type

* convert : fix pre-tokenizer type writing

* lint : fix

* make : add test-tokenizer-0-llama-v3

* wip

* models : add llama v3 vocab file

* llama : adapt punctuation regex + add llama 3 regex

* minor

* unicode : set bomb

* unicode : set bomb

* unicode : always use std::wregex

* unicode : support \p{N}, \p{L} and \p{P} natively

* unicode : try fix windows

* unicode : category support via std::regex

* unicode : clean-up

* unicode : simplify

* convert : add convert-hf-to-gguf-update.py

ggml-ci

* lint : update

* convert : add falcon

ggml-ci

* unicode : normalize signatures

* lint : fix

* lint : fix

* convert : remove unused functions

* convert : add comments

* convert : exercise contractions

ggml-ci

* lint : fix

* cmake : refactor test targets

* tests : refactor vocab tests

ggml-ci

* tests : add more vocabs and tests

ggml-ci

* unicode : cleanup

* scripts : ignore new update script in check-requirements.sh

* models : add phi-3, mpt, gpt-2, starcoder

* tests : disable obsolete

ggml-ci

* tests : use faster bpe test

ggml-ci

* llama : more prominent warning for old BPE models

* tests : disable test-tokenizer-1-bpe due to slowness

ggml-ci

---------

Co-authored-by: Jaggzh <jaggz.h@gmail.com>
Co-authored-by: Kazim Abrar Mahi <kazimabrarmahi135@gmail.com>
2024-04-29 16:58:41 +03:00
Tristan Druyen
abd3314064
llama : add phi 3 chat template (#6857)
* Add phi 3 chat template & tests

* test : fix chat template result

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-04-24 11:52:37 +03:00
Wouter
7dbdba5690
llama : add llama-3 chat template (#6751)
* Added llama-3 chat template

* Update llama.cpp

Co-authored-by: Samuel Tallet <36248671+SamuelTallet@users.noreply.github.com>

* Update llama.cpp

Co-authored-by: Samuel Tallet <36248671+SamuelTallet@users.noreply.github.com>

* Update tests/test-chat-template.cpp

Co-authored-by: Samuel Tallet <36248671+SamuelTallet@users.noreply.github.com>

* Added EOS stop sequence according to https://github.com/ggerganov/llama.cpp/pull/6751#issuecomment-2065602862

* Removed adding of BOS token before first message

* Removed bos token from expected output from llama-3

* Update tests/test-chat-template.cpp

Co-authored-by: Rene Leonhardt <65483435+reneleonhardt@users.noreply.github.com>

* Update tests/test-chat-template.cpp

Co-authored-by: Rene Leonhardt <65483435+reneleonhardt@users.noreply.github.com>

* Added <|end_of_text|> as another stop token

* Reverted last change of adding the end_of_text stop word for llama 3

---------

Co-authored-by: Wouter Tichelaar <tichelaarw@spar.net>
Co-authored-by: Samuel Tallet <36248671+SamuelTallet@users.noreply.github.com>
Co-authored-by: Rene Leonhardt <65483435+reneleonhardt@users.noreply.github.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-04-21 16:03:39 +03:00
slaren
0d56246f4b
ggml : group all experts in a single ggml_mul_mat_id (#6505)
* ggml : group all experts in a single ggml_mul_mat_id
cuda : improve mmid row copy

* cuda : fix bin bcast with non-cont src0

* test-backend-ops : only run all mul mat tests for base types

* llama : disable moe offloading with SYCL

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-04-18 15:18:48 +02:00
Shijie
f4dea7da18
llama : add qwen2moe (#6074)
* support qwen2moe

* fix-review

* metal : support unary ops for nelements % 4 != 0

* metal : require contiguousness for float4 unary kernels

* metal : require contiguousness for float4 unary kernels (cont)

* fix-review

* names : for brevity "SHARED_EXP" -> "SHEXP"

* llama : reuse build_moe_ffn()

* llama : add model type name

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-04-16 18:40:48 +03:00
Olivier Chafik
7593639ce3
main: add --json-schema / -j flag (#6659)
* main: add --json-schema / -j

* json: move json-schema-to-grammar to common lib

* json: fix zig build
2024-04-15 18:35:21 +01:00
Chao Jiang
04fbc5f23e
Add Command R chat template (#6650)
* Add chat template for command-r model series

* Fix indentation

* Add chat template test for command-r models and update the implementation to trim whitespaces

* Remove debug print
2024-04-14 18:16:34 +02:00
Olivier Chafik
ab9a3240a9
JSON schema conversion: ️ faster repetitions, min/maxLength for strings, cap number length (#6555)
* json: rename python schema converter to make import easier

* server: skip null json_schema / grammar fields

* json: deps management for primitive rules (+ allow null values)

* json: optimize repetitions for minItems/maxItems and regexps: `a{,3}` goes from `"a"? "a"? "a"?` (explosive combos) to `(a (a (a)?)?)?`

* grammars: add troubleshooting section to readme

* json: cap length of numbers to 15 digits before/after decimal point

(avoids infinite gen, e.g. "one third" -> `0.333333333333...`)

* json: unify all repetition code (w/ or w/o sep)

* json: support string minLength/maxLength

* server+json: update server/README w/ result_format

* nits

* json: fix type error w/ python 3.8

* json: fix server/README (json_schema in /completion vs. result_format in /v1/chat/completions)

* json: simplify DOT `{"type": "string", "pattern": "^.$"}`

* json: remove recursion in opt_repetitions (avoids Python stack overflow)

* json: rm dead code

* json: rm useless assert & ggml.h import
2024-04-12 19:43:38 +01:00
slaren
fbbc030ba9
metal : unify mul_mv_id kernels (#6556) 2024-04-12 18:13:20 +02:00
Olivier Chafik
cbaadc9294
grammars: 1.5x faster inference w/ complex grammars (vector reserves / reuses) (#6609)
* grammars: reserve rejects & next candidates

* grammars: reuse new_stacks

* grammars: fix missing sig change in llama.h

* grammars: fix test (api changed)

* grammars: update gbnf-validator.cpp

* grammars: simpler syntax (no swap)
2024-04-11 19:47:34 +01:00
Clint Herron
57dd02c44b
Tests: Added integration tests for GBNF parser (#6472)
* Added integration tests for GBNF parser to validate correctness of parsing, as well as correctness of string matching. Intended for use to pin behavior while working on performance improvements.

* Fixing whitespace errors and cleaning error message alert to be clearer.

* Removing hacky include to llama.cpp from grammar integration test now that needed functions are available via internal API.

* Comment cleanup.

* Reorganizing tests for readability.

* Cleaning up debug message to make a bit more sense.
2024-04-06 10:31:33 -04:00
kaizau
1ff4d9f3d6
Add OpenChat, Alpaca, Vicuna chat templates (#6397)
* Add openchat chat template

* Add chat template test for openchat

* Add chat template for vicuna

* Add chat template for orca-vicuna

* Add EOS for vicuna templates

* Combine vicuna chat templates

* Add tests for openchat and vicuna chat templates

* Add chat template for alpaca

* Add separate template name for vicuna-orca

* Remove alpaca, match deepseek with jinja output

* Regenerate chat template test with add_generation_prompt

* Separate deepseek bos from system message

* Match openchat template with jinja output

* Remove BOS token from templates, unprefix openchat
2024-04-03 17:24:31 +02:00
slaren
08a0c02060
ggml : mul_mat_id use the same tensor for all the experts (#6387)
* ggml : update mul_mat_id to use the same tensor for all the experts

* update cuda

* minor

* update metal

* update test-backend-ops

* fix cuda

* Update ggml-metal.m

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* update convert.py

* update convert-hf-to-gguf.py

* update convert.py for mixtral hf models

* Update convert-hf-to-gguf.py

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* cuda : support non-pow-2 number of experts

* allow quantize to work for split and merged experts models in the same way

* cleanup + disable mmap automatically with split tensors models

* update imatrix

* test-backend-ops : test qwen argsort

* update grok model loading

* llama : add merged experts tensors to the grok tensor map

* minor

* gguf : bump version

* fix quantizing of merged experts

* convert-hf-to-gguf.py : update grok (untested)

* make linter happy

* cuda/argsort : use shared memory instead of pool memory

* convert : fix grok tensor names

* metal : add support for non-pow-2 argsort

* llama : more loader cleanup, better error checking

* cuda : fix warning

* llama : still use mmap for loading old models, but copy the data to a host buffer

* add review note

* llama : remove ffn tensor counting + add sanity check

ggml-ci

* convert : fix handling of n_experts == None

ggml-ci

* imatrix : fix ncall counters

* llama : produce error if imatrix size does not match

* quantize : terminate on errors + trace logs

ggml-ci

* metal : pad shared memory to 16 bytes

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-04-03 16:07:05 +03:00
Kawrakow
55c1b2a3bb
IQ1_M: 1.75 bpw quantization (#6302)
* iq1_m: basics

* iq1_m: basics-2

* iq1_m: CUDA dequantize works

Very 1st shot I get PPL = 9.76 for LLaMA-v2-7B.

* iq1_m: separate shifts for each group of 8 in a block

We get
PPL(LLaMA-v2-7B ) = 9.2810
PPL(LLaMA-v2-13B) = 6.8105

Not bad, but slightly higher than
  sqrt(PPL(IQ1_S) * PPL(IQ2_XXS))
which is the expected outcome given that IQ1_M is
halfway between IQ1_S and IQ2_XXS in terms of bpw.
From this, we would expect
 PPL = 9.14 for LLaMA-v2-7B
 PPL = 6.63 for LLaMA-v2-13B

* iq1_m: go to 3-bit scales

There is slight increase in PPL, but the 0.0625 bpw reduction
in size is totally worth it.

We now have
PPL(LLaMA-v2-7B ) = 9.4469 at 1.96 bpw
PPL(LLaMA-v2-13B) = 6.8717 at 1.93 bpw
PPL(LLaMA-v2-70B) = 4.8568 at 1.85 bpw

* iq1_m: scalar dot product

* iq1_m: AVX2 dot product

* iq1_m: very slightly faster AVX2 dot product

* iq1_m: ARM_NEON dot product

Works, but very slow (10.5 t/s)

* iq1_m: Metal - dequantize works, dot product does not

* iq1_m: Metal now works

About the same performance as iq1_s.

* iq1_m: minor

* iq1_m: checking pure iq1_m quantization

It is pretty bad: PPL(LLaMA-v2-7B) = 34 if we quantize output.weight
with Q4_K.

* iiq1_m: slightly faster ARM_NEON dot product

10.5 t/s -> 11.65 t/s

* iq1_m: faster ARM_NEON dot product

11.65 t/s -> 14.9 t/s

* iq1_m: another minor ARM_NEON dot product improvement

14.9 -> 15.0 t/s

* iq1_m: small PPL improvement via super-block scale adjustment

After quantizing block scales redo the super-block scale fit.

PPL(LLaMA-v2-7B ) = 9.3346
PPL(LLaMA-v2-13B) = 6.8419
PPL(LLaMA-v2-70B) = 4.8294
PPL(Mistral-7B  ) = 8.1624

* iq1_m: adapt to CUDA refactoring

* iq1_m: remove unused variable

We have progressed to warnings being errors.

* iq1_m: add to backend-ops tests

* iq1_m: fix Windows ARM

* iq1_m: use common definition of iq1m_scale_t

* cuda: assert -> NO_DEVICE_CODE

* iq1_M: PR comments

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
2024-03-26 15:21:27 +01:00
Kawrakow
1f2fd4e727
tests : include IQ2_XXS and IQ2_XS in test-quantize-fns (#6303)
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
2024-03-25 19:33:15 +02:00
Olivier Chafik
f77a8ffd3b
tests : conditional python & node json schema tests (#6207)
* json: only attempt python & node schema conversion tests if their bins are present

Tests introduced in https://github.com/ggerganov/llama.cpp/pull/5978
disabled in https://github.com/ggerganov/llama.cpp/pull/6198

* json: orange warnings when tests skipped

* json: ensure py/js schema conv tested on ubuntu-focal-make

* json: print env vars in test
2024-03-22 15:09:07 +02:00
Olivier Chafik
72114edf06
json-schema-to-grammar : fix order of props + non-str const/enum (#6232)
* json: ordered json in server/schema converter to respect orig order

* json: ws nits

* json: support non-string const / enums
2024-03-22 15:07:44 +02:00
Georgi Gerganov
95d576b48e
metal : pad n_ctx by 32 (#6177)
* metal : require ne00 >= 128 for mat-mat kernels

ggml-ci

* llama : pad n_ctx by 32

ggml-ci
2024-03-22 09:36:03 +02:00
Georgi Gerganov
924ce1dce7
tests : disable system() calls (#6198)
ggml-ci
2024-03-21 16:20:05 +02:00
Olivier Chafik
5b7b0ac8df
json-schema-to-grammar improvements (+ added to server) (#5978)
* json: fix arrays (disallow `[,1]`)

* json: support tuple types (`[number, string]`)

* json: support additionalProperties (`{[k: string]: [string,number][]}`)

* json: support required / optional properties

* json: add support for pattern

* json: resolve $ref (and support https schema urls)

* json: fix $ref resolution

* join: support union types (mostly for nullable types I think)

* json: support allOf + nested anyOf

* json: support any (`{}` or `{type: object}`)

* json: fix merge

* json: temp fix for escapes

* json: spaces in output and unrestricted output spaces

* json: add typings

* json:fix typo

* Create ts-type-to-grammar.sh

* json: fix _format_literal (json.dumps already escapes quotes)

* json: merge lit sequences and handle negatives

{"type": "string", "pattern": "^({\"question\": \"[^\"]+\", \"response\": \"[^\"]+\"}\\n)+$"}

* json: handle pattern repetitions

* Update json-schema-to-grammar.mjs

* Create regex-to-grammar.py

* json: extract repeated regexp patterns to subrule

* Update json-schema-to-grammar.py

* Update json-schema-to-grammar.py

* Update json-schema-to-grammar.py

* json: handle schema from pydantic Optional fields

* Update json-schema-to-grammar.py

* Update json-schema-to-grammar.py

* Update ts-type-to-grammar.sh

* Update ts-type-to-grammar.sh

* json: simplify nullable fields handling

* json: accept duplicate identical rules

* json: revert space to 1 at most

* json: reuse regexp pattern subrules

* json: handle uuid string format

* json: fix literal escapes

* json: add --allow-fetch

* json: simplify range escapes

* json: support negative ranges in patterns

* Delete commit.txt

* json: custom regex parser, adds dot support & JS-portable

* json: rm trailing spaces

* Update json-schema-to-grammar.mjs

* json: updated server & chat `( cd examples/server && ./deps.sh )`

* json: port fixes from mjs to python

* Update ts-type-to-grammar.sh

* json: support prefixItems alongside array items

* json: add date format + fix uuid

* json: add date, time, date-time formats

* json: preserve order of props from TS defs

* json: port schema converter to C++, wire in ./server

* json: nits

* Update json-schema-to-grammar.cpp

* Update json-schema-to-grammar.cpp

* Update json-schema-to-grammar.cpp

* json: fix mjs implementation + align outputs

* Update json-schema-to-grammar.mjs.hpp

* json: test C++, JS & Python versions

* json: nits + regen deps

* json: cleanup test

* json: revert from c++17 to 11

* json: nit fixes

* json: dirty include for test

* json: fix zig build

* json: pass static command to std::system in tests (fixed temp files)

* json: fix top-level $refs

* json: don't use c++20 designated initializers

* nit

* json: basic support for reserved names `{number:{number:{root:number}}}`

* Revamp test cmake to allow args (WORKING_DIRECTORY needed for JSON test)

* json: re-ran server deps.sh

* json: simplify test

* json: support mix of additional props & required/optional

* json: add tests for some expected failures

* json: fix type=const in c++, add failure expectations for non-str const&enum

* json: test (& simplify output of) empty schema

* json: check parsing in test + fix value & string refs

* json: add server tests for OAI JSON response_format

* json: test/fix top-level anyOf

* json: improve grammar parsing failures

* json: test/fix additional props corner cases

* json: fix string patterns (was missing quotes)

* json: ws nit

* json: fix json handling in server when there's no response_format

* json: catch schema conversion errors in server

* json: don't complain about unknown format type in server if unset

* json: cleaner build of test

* json: create examples/json-schema-pydantic-example.py

* json: fix date pattern

* json: move json.hpp & json-schema-to-grammar.{cpp,h} to common

* json: indent 4 spaces

* json: fix naming of top-level c++ function (+ drop unused one)

* json: avoid using namespace std

* json: fix zig build

* Update server.feature

* json: iostream -> fprintf

* json: space before & refs for consistency

* json: nits
2024-03-21 11:50:43 +00:00
Xuan Son Nguyen
aab606a11f
llama : add Orion chat template (#6066) 2024-03-15 10:44:57 +02:00
slaren
d8fd0ccf6a
test-backend-ops : skip CPU backend by default (#6028) 2024-03-13 15:58:30 +02:00
Georgi Gerganov
83796e62bc
llama : refactor unicode stuff (#5992)
* llama : refactor unicode stuff

ggml-ci

* unicode : names

* make : fix c++ compiler

* unicode : names

* unicode : straighten tables

* zig : fix build

* unicode : put nfd normalization behind API

ggml-ci

* swift : fix build

* unicode : add BOM

* unicode : add <cstdint>

ggml-ci

* unicode : pass as cpts as const ref
2024-03-11 17:47:47 +02:00
Georgi Gerganov
5b09797321
ggml : remove old quantization functions (#5942)
* ggml : remove old quantization functions

ggml-ci

* ggml : simplify ggml_quantize_chunk

ggml-ci

* ggml : restrict correctness

ggml-ci

* ggml : remove hist data from the quantization API

ggml-ci

* tests : remove hist usage in test-backend-ops

ggml-ci

* vulkan : remove hist and fix typo
2024-03-09 15:53:59 +02:00
Georgi Gerganov
2c4f566c88
tests : gitignore ggml-common.h 2024-03-09 14:17:11 +02:00
leejet
7d43c585dc
add some new ops, fix some operators and add batch operations to certain operators. (ggml/747)
* cuda: fix group_norm

* cuda: add batch inference support for ggml_pad/ggml_upscale

* add ggml_arrange

* add ggml_timestep_embedding

* update ggml_arange/ggml_timestep_embedding tests

* cuda: fix im2col

* add ggml_arange/ggml_timestep_embbeding support for metal backend

* fix some bugs

* fix some bugs

* Update ggml.h

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update ggml-cuda.cu

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update ggml-metal.m

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update ggml-metal.m

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update ggml-metal.metal

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* modify according to the review comments

* ggml : fix compile warnings + code style

* ggml : normalize compute_forward calls + fix seg fault in debug

* minor

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
2024-03-04 10:39:10 +02:00
Kawrakow
0becb22ac0
IQ4_XS: a 4.25 bpw quantization (#5747)
* Try IQ4_NL with blocks of 64 - does not look good

* iq4_xs: go to super-blocks of 256 and 6-bit scales for blocks of 32

* iq4_xs: CUDA works - 133.2 t/s

* iq4_xs: AVX2 dot product

* iq4_xs: ARM_NEON dot product

* iq4_nl: Metal implementation

As usual, Metal / Apple Silicon don't like my quants.

* iq3_xs: minor fix

* iq4_xs: shrink by using IQ3_S for attn_k and attn_q

* iq4_xs: revert using IQ3_S for attn_k and attn_v

PPL vs size is good, but CPU performance suffers: on M2 Max
TG-128 drops to 21.7 t/s from 28.8, and on a Ryzen-7950X
to 14.5 t/s from 15.8 t/s. On CUDA we have 135 t/s when
using IQ3_S vs 133 t/s with pure IQ4_XS.

* Fix CI

* iq4_xs: Added forgotten check for 256 divisibility

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
2024-02-27 16:34:24 +02:00
Kawrakow
a33e6a0d2a
Adding IQ2_S and IQ2_M to complete coverage of the 2-3 bit quantization range (#5721)
* Adding IQ2_S and IQ2_M as a single cumulative commit

* Update examples/quantize/quantize.cpp

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-02-26 18:28:38 +02:00
Georgi Gerganov
ab336a9d5e
code : normalize enum names (#5697)
* coda : normalize enum names

ggml-ci

* code : cont

* code : cont
2024-02-25 12:09:09 +02:00
Kawrakow
4c4cb30736
IQ3_S: a much better alternative to Q3_K (#5676)
* iq4_nl: squash commits for easier rebase

* Basics (quantize, dequantize)
* CUDA dequantize and dot product
* Slightly faster CUDA dot product (120 t/s)
* Switch to 6-bit scales
* Scalar dot product
* AVX2 dot product
* ARM_NEON dot product
* Works on metal, but still slow
* Slightly better Metal dot product
* Another small Metal improvement
* Metal dot product is getting there
* Faster CUDA dot product
* Add 1/8 ffn_down layers as Q5_K when no imatrix has been provided
* Report the actual bpw
* Add _xs mix that is 4.05 bpw for non-MoE models
* Remove IQ4_XS for now, slightly adjust kvalues_iq4nl
* AVX2 dot product uses Q8_0 instead of Q8_K
* Add to test-backend-ops
* Minor fix
* Also use use Q5_K for attn_output in MoE models
* Fixes after merging latest master
* Switching to blocks of 32
* AVX2 for blocks of 32
* Scaler dot product for blocks of 32
* ARM_NEON dot product for blocks of 32
* Metal kernels for blocks of 32
* Slightly faster Metal kernels

* Resurrecting iq3_xs

After all the experimentation, nothing was better than this.

* Minor PPL improvement via a block scale fudge factor

* Minor improvement via 3 neighbours

* iq3_xs: working scalar and AVX2 dot products

* iq3_xs: ARM_NEON dot product - works but extremely slow (10 t/s)

* iq3_xs: working Metal implementation

* Adding IQ3_M - IQ3_XS mix with mostly Q4_K

* iiq3_xs: a 3.4375 bpw variant

* iq3_xs: make CUDA work for new version

* iq3_xs: make scalar and AVX2 work for new version

* iq3_s: make ARM_NEON work with new version

* iq3_xs: make new version work on metal

Performance is very similar to Q3_K_S

* iq3_xs: tiny Metal speed improvement

* iq3_xs: tiny Metal speed improvement

* Fix stupid warning

* Q3_K_XS now uses a mix of IQ3_XS and IQ3_XXS

* iq3_xs: rename to iq3_s

* iq3_s: make tests pass

* Move Q3_K_XS mix to 3.25 bpw

* Attempt to fix failing tests

* Another attempt to fix the Windows builds

* Attempt to fix ROCm

* ROCm again

* iq3_s: partial fix for QK_K = 64

* iq3_s: make it work on metal for QK_K = 64

Pleasent surprise: the coding was super-block size independent,
so all it took was to delete some QK_K == 256 guards.

* Will this fix ROCm?

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
2024-02-24 16:23:52 +02:00
Xuan Son Nguyen
373ee3fbba
Add Gemma chat template (#5665)
* add gemma chat template

* gemma: only apply system_prompt on non-model message
2024-02-22 19:10:21 +01:00
Xuan Son Nguyen
a46f50747b
server : fallback to chatml, add AlphaMonarch chat template (#5628)
* server: fallback to chatml

* add new chat template

* server: add AlphaMonarch to test chat template

* server: only check model template if there is no custom tmpl

* remove TODO
2024-02-22 10:33:24 +02:00
Kawrakow
a14679cc30
IQ4_NL: 4-bit non-linear quants with blocks of 32 (#5590)
* iq4_nl: squash commits for easier rebase

* Basics (quantize, dequantize)
* CUDA dequantize and dot product
* Slightly faster CUDA dot product (120 t/s)
* Switch to 6-bit scales
* Scalar dot product
* AVX2 dot product
* ARM_NEON dot product
* Works on metal, but still slow
* Slightly better Metal dot product
* Another small Metal improvement
* Metal dot product is getting there
* Faster CUDA dot product
* Add 1/8 ffn_down layers as Q5_K when no imatrix has been provided
* Report the actual bpw
* Add _xs mix that is 4.05 bpw for non-MoE models
* Remove IQ4_XS for now, slightly adjust kvalues_iq4nl
* AVX2 dot product uses Q8_0 instead of Q8_K
* Add to test-backend-ops
* Minor fix
* Also use use Q5_K for attn_output in MoE models
* Fixes after merging latest master
* Switching to blocks of 32
* AVX2 for blocks of 32
* Scaler dot product for blocks of 32
* ARM_NEON dot product for blocks of 32
* Metal kernels for blocks of 32
* Slightly faster Metal kernels

* iq4_nl: Fix after merging with master

* iq4_nl: another fix after merging with master

* Use IQ4_NL instead of Q4_K when using k-quants is not possible

* Fix typo that makes several tests fail

* It was the ggml_vdotq thing missed inside the brackets

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
2024-02-21 11:39:52 +02:00
Xuan Son Nguyen
11b12de39b
llama : add llama_chat_apply_template() (#5538)
* llama: add llama_chat_apply_template

* test-chat-template: remove dedundant vector

* chat_template: do not use std::string for buffer

* add clarification for llama_chat_apply_template

* llama_chat_apply_template: add zephyr template

* llama_chat_apply_template: correct docs

* llama_chat_apply_template: use term "chat" everywhere

* llama_chat_apply_template: change variable name to "tmpl"
2024-02-19 10:23:37 +02:00
Herman Semenov
5d3de51f97
ggml, common, examples, tests : fixed type arguments in printf (#5528) 2024-02-18 18:20:12 +02:00
Kawrakow
bd2d4e393b
1.5 bit quantization (#5453)
* iq1_s: WIP basics

* iq1_s: CUDA is working

* iq1_s: scalar CPU dot product

* iq1_s: WIP AVX2 dot product - something is not right

* Fix tests

* Fix shadow warnings

* Fix after merge with latest master

* iq1_s: AVX2 finally works

* iq1_s: ARM_NEON dot product. Works, but not very fast

* iq1_s: better grid

* iq1_s: use IQ2_XXS for attn_output

At a cost of 0.04 extra bpw this gives a big improvement in PPL.

* iq1_s: Metal basics

Dequantize works, but not dot product

* iq1_s: Metal works, but quite slow

As usual, Apple Silicon does not like the code I write.

* iq1_s: Tests

* iq1_s: slightly faster dot product

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
2024-02-18 18:16:55 +02:00
Georgi Gerganov
8f1be0d42f
ggml : add ALiBi support for ggml_soft_max_ext (#5488)
* ggml : avoid recomputing alibi slopes (CPU)

* llama : reuse hparams.f_max_alibi_bias in all cases

ggml-ci

* ggml : support alibi bias in ggml_soft_max_ext (CPU + Metal)

ggml-ci

* ggml : handle all SRCs (do not break on first null)

ggml-ci

* tests : do not use slope for large soft_max

accumulates too much error

ggml-ci

* ggml : alternative ALiBi without extra tensor

We compute the slopes in the kernel

ggml-ci

* cuda : add ALiBi support in ggml_soft_max_ext

ggml-ci

* ggml : deprecate ggml_alibi

* ggml : support multi-sequence ALiBi (Metal)

ggml-ci

* cuda : add multi-seq ALiBi + remote F16 soft_max

ggml-ci

* ggml : update deprecation message

* ggml : fix pos ptr when no ALiBi

ggml-ci

* cuda : fix performance (pow -> powf)

* cuda : precompute ALiBi constants

* metal : pre-compute ALiBi slopes

ggml-ci

* llama : init kq_pos only if needed

ggml-ci

* test-backend-ops : add null pos test to soft_max

test-backend-ops : replace soft_max tests

ggml-ci

---------

Co-authored-by: slaren <slarengh@gmail.com>
2024-02-17 23:04:16 +02:00
bmwl
f486f6e1e5
ggml : add numa options (#5377)
* Added numa options to allow finer grained control as well as plumbing for a new mirror mode that will require numa.h

* Reverted Makefile

* Fixed include

* Removed sched.h from ggml.h, moved ggml_get_numa_affinity into ggml.c, removed trailing whitespace and fixed up a few inconsistent variables

* removed trailing whitespace

* Added numa options to allow finer grained control as well as plumbing for a new mirror mode that will require numa.h

* Reverting Makefile

* Fixed a number of issues with the move from BOOL to ggml_numa_strategies. Added a note about mirror mode note being implemented yet

* Removing MIRROR_MODE code for this PR

* Removing last bit of MIRROR_MODE code for this PR

* Removing unneeded branch in server.cpp example and moving get_numa_affinity and making it static

* Fixed lingering init_llama_backend() bool calls in tests and examples

* Remote enum llama_numa_strategies

* Revert bad merge with dynatemp flags

* add missing enum ggml_numa_strategies declaration and revert sync problem with master

* add missing enum ggml_numa_strategies declaration

* fixed ggml_init_numa variable

* Update ggml.h

Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>

* Update READMEs with info about numa flags, change INTERLEAVE strategy name to DISTRIBUTE everywhere, implement the improved distribution strategy from @rankaiyx, fix a spelling mistake and un-merge some bad merges

* split numa init out from llama_backend_init and created llama_numa_init. Updated all code paths and samples

* Fix up some boolean vs enum comparisons

* Added #ifdefs for non-Linux OS that don't have cpu_set_t datatype

* Update ggml.h

Align enum values

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update ggml.c

Remove whitespace

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update ggml.c

align paremeters

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update examples/server/server.cpp

remove whitespace and align brace

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update common/common.cpp

Remove whitespace and align brace

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* unified ggml_numa_strategy enum and fixed text alignment in server.cpp example

* Update ggml.c

simplified return for platforms without NUMA support

Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>

* removed redundant else from cli argument processing of --numa

* whitespace

---------

Co-authored-by: root <root@nenya.lothlorien.ca>
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Jared Van Bortel <jared@nomic.ai>
2024-02-16 11:31:07 +02:00
Georgi Gerganov
cf45252a7c
tests : multi-thread the tokenizer tests (#5474)
* tests : multi-thread the tokenizer tests

ggml-ci

* unicode : fix data race for unidentified codepoints

ggml-ci

* unicode : minor style fixes

ggml-ci
2024-02-13 15:14:22 +02:00
Georgi Gerganov
99b8b43d7b
tests : disable moe test (#5473) 2024-02-13 11:20:24 +02:00
snadampal
a07d0fee1f
ggml : add mmla kernels for quantized GEMM (#4966)
* ggml: aarch64: implement smmla kernel for q8_0_q8_0 quantized gemm

armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q8_0_q8_0 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"

On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.

* ggml: aarch64: implement smmla kernel for q4_0_q8_0 quantized gemm

armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q4_0_q8_0 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"

On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.

* ggml: aarch64: implement smmla kernel for q4_1_q8_1 quantized gemm

armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q4_1_q8_1 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"

On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.

* ggml: update unit tests for the new vec_dot interface

* llama.cpp: add MATMUL_INT8 capability to system_info
2024-02-11 15:22:33 +02:00
Johannes Gäßler
26d4efd11e
sampling: fix top_k <= 0 (#5388)
* sampling: fix top_k <= 0

* Update llama.cpp

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-02-08 09:46:30 +01:00
Georgi Gerganov
8504d2d0da
tests : .gitignore obj files 2024-02-08 09:46:47 +02:00
Michael Klimenko
52bb63c708
refactor : switch to emplace_back to avoid extra object (#5291) 2024-02-03 13:23:37 +02:00
JidongZhang-THU
15606309a0
llava : add MobileVLM support (#5132)
* New Feature:
    1. Sum_Rows:
        fix cuda kernel overflow
        fix block shape error when nrows too big
    2. Im2Col:
        Support Batch in cuda
        Support f32 to f32 both in cpu && cuda
    3. DepthWiseConv:
        Support by Im2Col && MulMat
    4. Pool_2d:
        Supoort avg pooling in cuda
    5. HardSigmoid:
        Imp in cuda
    6. HardSwish:
        Imp in cuda

* fix tabs instead of spaces

* code clean

* CUDA POOL2D

* ADD POOL2D test case in test-backend-ops.cpp

* code clean

* fix pool2d_kernel

nits

* fix bug in pool2d kernel

* fix avg pooling, count_include_pad

nits

* test-backend-ops : add more pool_2d tests

* cuda : fix warnings and formatting

* ggml : check types in release builds too in pool_2d

* test-backend-ops : remove f16 pool_2d tests

* cuda : more style fixes

* Add assert in ggml_cuda_op_pool2d

* pool2d float padding fallback

* test-backend-ops : add dst_type to im2col

---------

Co-authored-by: slaren <slarengh@gmail.com>
2024-01-31 15:10:15 +02:00
John Balis
625a699b54
ggml_cuda_cpy support for 4d tensors and float16->float32 upcasting (ggml/686)
* added cuda float16->float32 upcasting to ggml_cuda_cpy

* added ability to copy 4d tensors with the cuda backend

* added tests for float16_>float32 upcast and 4d tensor cuda copys

* added 4d copy test for float32->float16 copy

* applied patch suggested by @iamlemec

* simplify cpy tests

---------

Co-authored-by: slaren <slarengh@gmail.com>
2024-01-30 16:20:25 +02:00
Kawrakow
f4d7e54974
SOTA 3-bit quants (#5196)
* iq3_xxs: quantize/dequantize

RMSE seems a bit high-ish at about half-way between q2_K and
q3_K, so need to check more.

* iq3_xxs: CUDA dequantize works

* iq2_xxs: tuning quantization

* iq3_xxs: starting to look better

PPL on wiki.test.raw
LLaMA-v1-7B: 6.4218
LLaMA-v2-7B: 6.3560
Mistral-7B : 6.0717

This is better than Q3_K_XS, with a 5% reduction in quantized model
size.

* iq3_xxs: CUDA dot product

We have
PP-512: 5891 t/s
TG-128: 143.9 t/s

* iq3_xxs: scalar and AVX2 dot products

* iq3_xxs: ARM_NEON and Metal

Metal performance is decent, ARM_NEON is pathetic

* iq3_xxs: slightly better grid points

* Faster iq3_xxs and iq2_xs dot products on CUDA

* iq3_xxs: add some quant mix

* iq3_xxs: fix failing quantization test

Dot product still fails. Is this real?

* iq3_xxs: hopefully fix ROCm

* iq3_xxs: failing tests

This time the dot product accuracy did find an actual bug
in the AVX2 implementation.

* Add IQ3_XXS to test-backend-ops

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
2024-01-30 15:14:12 +02:00
Jared Van Bortel
fbf1ddec69
Nomic Vulkan backend (#4456)
Signed-off-by: Jared Van Bortel <jared@nomic.ai>
Co-authored-by: niansa <anton-sa@web.de>
Co-authored-by: Adam Treat <treat.adam@gmail.com>
Co-authored-by: Aaron Miller <apage43@ninjawhale.com>
Co-authored-by: ToKiNoBug <tokinobug@163.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
2024-01-29 15:50:50 -05:00
Abhilash Majumder
0f648573dd
ggml : add unified SYCL backend for Intel GPUs (#2690)
* first update for migration

* update init_cublas

* add debug functio, commit all help code

* step 1

* step 2

* step3 add fp16, slower 31->28

* add GGML_LIST_DEVICE function

* step 5 format device and print

* step6, enhance error check, remove CUDA macro, enhance device id to fix none-zero id issue

* support main device is non-zero

* step7 add debug for code path, rm log

* step 8, rename all macro & func from cuda by sycl

* fix error of select non-zero device, format device list

* ren ggml-sycl.hpp -> ggml-sycl.h

* clear CMAKE to rm unused lib and options

* correct queue: rm dtct:get_queue

* add print tensor function to debug

* fix error: wrong result in 658746bb26702e50f2c59c0e4ada8e9da6010481

* summary dpct definition in one header file to replace folder:dpct

* refactor device log

* mv dpct definition from folder dpct to ggml-sycl.h

* update readme, refactor build script

* fix build with sycl

* set nthread=1 when sycl, increase performance

* add run script, comment debug code

* add ls-sycl-device tool

* add ls-sycl-device, rm unused files

* rm rear space

* dos2unix

* Update README_sycl.md

* fix return type

* remove sycl version from include path

* restore rm code to fix hang issue

* add syc and link for sycl readme

* rm original sycl code before refactor

* fix code err

* add know issue for pvc hang issue

* enable SYCL_F16 support

* align pr4766

* check for sycl blas, better performance

* cleanup 1

* remove extra endif

* add build&run script, clean CMakefile, update guide by review comments

* rename macro to intel hardware

* editor config format

* format fixes

* format fixes

* editor format fix

* Remove unused headers

* skip build sycl tool for other code path

* replace tab by space

* fix blas matmul function

* fix mac build

* restore hip dependency

* fix conflict

* ren as review comments

* mv internal function to .cpp file

* export funciton print_sycl_devices(), mv class dpct definition to source file

* update CI/action for sycl code, fix CI error of repeat/dup

* fix action ID format issue

* rm unused strategy

* enable llama_f16 in ci

* fix conflict

* fix build break on MacOS, due to CI of MacOS depend on external ggml, instead of internal ggml

* fix ci cases for unsupported data type

* revert unrelated changed in cuda cmake
remove useless nommq
fix typo of GGML_USE_CLBLAS_SYCL

* revert hip cmake changes

* fix indent

* add prefix in func name

* revert no mmq

* rm cpu blas duplicate

* fix no_new_line

* fix src1->type==F16 bug.

* pass batch offset for F16 src1

* fix batch error

* fix wrong code

* revert sycl checking in test-sampling

* pass void as arguments of ggml_backend_sycl_print_sycl_devices

* remove extra blank line in test-sampling

* revert setting n_threads in sycl

* implement std::isinf for icpx with fast math.

* Update ci/run.sh

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update examples/sycl/run-llama2.sh

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update examples/sycl/run-llama2.sh

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update CMakeLists.txt

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update CMakeLists.txt

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update CMakeLists.txt

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update CMakeLists.txt

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* add copyright and MIT license declare

* update the cmd example

---------

Co-authored-by: jianyuzh <jianyu.zhang@intel.com>
Co-authored-by: luoyu-intel <yu.luo@intel.com>
Co-authored-by: Meng, Hengyu <hengyu.meng@intel.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-01-28 17:56:23 +02:00
Johannes Gäßler
b2b2bf988c
Tests for min_p, sampling queue (#5147) 2024-01-28 09:35:14 +01:00
Michael Klimenko
35a2ee9143
Remove unused data and add fixes (#5154)
* Remove unused data and add fixes

* Add missing file

* Address review comments

* Replace the scope of vq allocation
2024-01-27 15:25:55 +01:00
Georgi Gerganov
3b7c914de2
tests : gitignore test-c.o 2024-01-26 14:48:15 +02:00
crasm
413e7b0559
ci : add model tests + script wrapper (#4586)
* scripts : add lib.sh and lib_test.sh

* scripts : stub out new ci-run.sh script

* scripts : switch to PascalCase for functions

This looks a little odd at first, but I find it very useful as a
convention to know if a command is part of our code vs a builtin.

* scripts : add some fancy conversion from snake_case to PascalCase

* Add venv to ci/run.sh

* Revert scripts work

* scripts : add wrapper script for local use of ci/run.sh

* Simplify .gitignore for tests, clang-tidy fixes

* Label all ctest tests

* ci : ctest uses -L main

* Attempt at writing ctest_with_model

* Update test-model-load-cancel

* ci : add ctest_with_model for debug and release

ggml-ci

* Fix gg_get_model function

ggml-ci

* got stuck on CMake

* Add get_model.cpp to tests/CMakeLists.txt

ggml-ci

* Fix README.md output for ctest_with_model

ggml-ci

* workflows : use `-L main` for all ctest

ggml-ci

* Fixes

* GG_RUN_CTEST_MODELFILE => LLAMACPP_TESTMODELFILE
* Always show warning rather than failing if model file variable is not
  set

* scripts : update usage text for ci-run.sh
2024-01-26 14:18:00 +02:00
Georgi Gerganov
38566680cd
ggml : add IQ2 to test-backend-ops + refactoring (#4990)
* ggml : add IQ2 to test-backend-ops + refactoring

ggml-ci

* cuda : update supports_op for IQ2

ggml-ci

* ci : enable LLAMA_CUBLAS=1 for CUDA nodes

ggml-ci

* cuda : fix out-of-bounds-access in `mul_mat_vec_q`

ggml-ci

* tests : avoid creating RNGs for each Q tensor

ggml-ci

* tests : avoid creating RNGs for each tensor

ggml-ci
2024-01-17 18:54:56 +02:00
Georgi Gerganov
c918fe8dca
metal : create autorelease pool during library build (#4970)
* metal : create autorelease pool during library build

ggml-ci

* test : simplify

ggml-ci
2024-01-17 18:38:39 +02:00
Kawrakow
147b17ac94
2-bit quantizations (#4897)
* imatrix: load

* imatrix: WIP

* imatrix: Add Q2_K quantization

* imatrix: also guard against Q2_K_S quantization without importance matrix

* imatrix: guard even more against low-bit quantization misuse

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
2024-01-14 09:45:56 +02:00
slaren
e7e4df031b
llama : ggml-backend integration (#4766)
* llama : ggml-backend integration

* ggml-backend : add names to buffers

* fix unmap after loading

* batched-bench : add tensor_split param

* llama : check for null tensor_split

* ggml-backend : increase GGML_MAX_BACKENDS

* improve graph splitting, partial fix for --no-kv-offload

* cuda : add ggml-backend split buffer support

* cuda : do not create buffer types for devices that don't exist (fixes usage without CUDA devices available)

* ggml : fix null backend dereference (#4807)

* ggml : fix null backend dereference

* ggml : also check ggml_backend_is_cpu

* test-backend-ops : check buffer allocation failures

* llama : add cparam (split_mode) and command line argument (--split-mode, -sm) to configure the split mode (none, layer or row)

* ggml : fix mul_mat_id work size

* llama : rewrite session kv load/set without graphs

* minor

* llama : only initialize used backends, free backends on context free

* llama : abort ctx if cuda backend init fails

* llama : rewrite lora with ggml-backend and compute on CPU

ggml-ci

* llama : only map to a backend buffer the region of the file mapping containing the tensors used in the buffer

* opencl : add ggml-backend buffer type

* cuda : only use batched_cublas with batched mat muls (fixes fp16 tg perf)

* llama : on Metal, by default offload the full model

ggml-ci

* metal : page align the data ptr (#4854)

* Apply suggestions from code review

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* cuda : fix split buffer free

* address review comments

* llama-bench : add split-mode parameter

* fix whitespace

* opencl : fix double initialization

* server : add --split-mode parameter

* use async copy and compute to improve multi-gpu performance

ggml-ci

* use async memcpys to copy the graph outputs to the CPU

* fix opencl

* use a host buffer for the cpu compute buffer for faster copies to the gpu

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
2024-01-12 20:07:38 +01:00
Kawrakow
49662cbed3
ggml : SOTA 2-bit quants (add IQ2_XS) (#4856)
* iq2_xs: basics

* iq2_xs: this should have been in the basics

* iq2_xs: CUDA and scalar CPU works

* iq2_xs: WIP Metal

* iq2_xs: Metal now works

* iq2_xs: working, but dog slow, ARM_NEON dot product

* iq2_xs: better ARM_NEON dot product

We are now at 19.5 t/s for TG-128 and 61 t/s for PP-512 when
running on the CPU.

* iq2_xs: AVX2 dot product - 19.5 t/s

* iq2_xs: faster AVX2 dit product

21.4 t/s for TG-128, 59.2 t/s for PP-512.
The latter is 2x compared to the previous version.

* iq2_xs: had forgotten to delete iq2-data.h

* Add llama enum for IQ2_XS

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
2024-01-11 21:39:39 +02:00
Johannes Gäßler
8f900abfc0
CUDA: faster softmax via shared memory + fp16 math (#4742) 2024-01-09 08:58:55 +01:00
Kawrakow
dd5ae06405
SOTA 2-bit quants (#4773)
* iq2_xxs: basics

* iq2_xxs: scalar and AVX2 dot products

Needed to change Q8_K to have quants in the -127...127 range,
else the IQ2_XXS AVX implementation becomes very awkward.
The alternative would have been to use Q8_0 instead. Perhaps
I'll change later, for now this is what we have.

* iq2_xxs: ARM_NEON dot product

Somehow strangely slow (112 ms/token).

* iq2_xxs: WIP Metal

Dequantize works, something is still wrong with the
dot product.

* iq2_xxs: Metal dot product now works

We have
PP-512 = 475 t/s
TG-128 = 47.3 t/s

Not the greatest performance, but not complete garbage either.

* iq2_xxs: slighty faster dot product

TG-128 is now 48.4 t/s

* iq2_xxs: slighty faster dot product

TG-128 is now 50.9 t/s

* iq2_xxs: even faster Metal dot product

TG-128 is now 54.1 t/s.

Strangely enough, putting the signs lookup table
into shared memory has a bigger impact than the
grid values being in shared memory.

* iq2_xxs: dequantize CUDA kernel - fix conflict with master

* iq2_xxs: quantized CUDA dot product (MMVQ)

We get TG-128 = 153.1 t/s

* iq2_xxs: slightly faster CUDA dot product

TG-128 is now at 155.1 t/s.

* iq2_xxs: add to llama ftype enum

* iq2_xxs: fix MoE on Metal

* Fix missing MMQ ops when on hipBLAS

I had put the ggml_supports_mmq call at the wrong place.

* Fix bug in qequantize_row_iq2_xxs

The 0.25f factor was missing.
Great detective work by @ggerganov!

* Fixing tests

* PR suggestion

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
2024-01-08 16:02:32 +01:00
Johannes Gäßler
a91928014f
Print backend name on test-backend-ops failure (#4751) 2024-01-04 09:43:23 +01:00
Guillaume Wenzek
5f66ebca9c ggml : extend ggml_get_rows, ggml_repeat, ggml_concat (ggml/639)
* add more int ops

* ggml_compute_forward_dup_bytes

* add tests

* PR comments

* tests : minor indentations

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-01-03 14:38:38 +02:00
Georgi Gerganov
58ba655af0
metal : enable shader debugging (cmake option) (#4705)
* ggml : disable fast-math for Metal (cmake build only)

ggml-ci

* metal : fix Metal API debug warnings

* cmake : add -fno-inline for Metal build (#4545)

* metal : fix API debug warnings

* metal : fix compile warnings

* metal : use uint64_t for strides

* cmake : rename option to LLAMA_METAL_SHADER_DEBUG

* metal : fix mat-vec Q8_0 kernel for BS > 1

* metal : normalize mat-vec kernel signatures

* cmake : respect LLAMA_QKK_64 option

* metal : fix mat-vec Q4_K kernel for QK_K == 64

ggml-ci
2024-01-02 10:57:44 +02:00
Cuong Trinh Manh
97bbca6e85
cmake : fix ld warning duplicate libraries libllama.a (#4671)
* fix "ld: warning: ignoring duplicate libraries: '../libllama.a'"

* fix warning in example.
2023-12-29 16:39:15 +02:00
bssrdf
afc8c19291
ggml : fix some mul mat cases + add tests for src1 F16 (ggml/669)
* fixed mul-mat error for old GPUs

* style fixes

* add mul mat src1 f16 test cases, fix more cases

ggml-ci

---------

Co-authored-by: bssrdf <bssrdf@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
2023-12-29 14:54:19 +02:00
manikbhandari
ea5497df5d
gpt2 : Add gpt2 architecture integration (#4555) 2023-12-28 15:03:57 +01:00
slaren
5bf3953d7e
cuda : improve cuda pool efficiency using virtual memory (#4606)
* cuda : improve cuda pool efficiency using virtual memory

* fix mixtral

* fix cmake build

* check for vmm support, disable for hip

ggml-ci

* fix hip build

* clarify granularity

* move all caps to g_device_caps

* refactor error checking

* add cuda_pool_alloc, refactor most pool allocations

ggml-ci

* fix hip build

* CUBLAS_TF32_TENSOR_OP_MATH is not a macro

* more hip crap

* llama : fix msvc warnings

* ggml : fix msvc warnings

* minor

* minor

* cuda : fallback to CPU on host buffer alloc fail

* Update ggml-cuda.cu

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* Update ggml-cuda.cu

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* ensure allocations are always aligned

* act_size -> actual_size

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
2023-12-24 14:34:22 +01:00
Georgi Gerganov
afefa319f1
ggml : change ggml_scale to take a float instead of tensor (#4573)
* ggml : change ggml_scale to take a float instead of tensor

* ggml : fix CPU implementation

* tests : fix test-grad0

ggml-ci
2023-12-21 23:20:49 +02:00
Ebey Abraham
b9e74f9bca
llama : add phi-2 + fix NeoX rope + ggml_mul_mat_set_prec (#4490)
* phi2 implementation

* fix breaking change

* phi-2 : various fixes

* phi-2 : use layer norm eps

* py : whitespaces

* llama : fix meta KV override bug

* convert : phi don't add BOS token

* convert : revert "added_tokens_decoder" change

* phi-2 : scale Q instead of KQ for better precision

* ggml : fix NeoX rope to rotate just first n_dims

* cuda : less diff in the rope_neox kernel

* ggml : add ggml_mul_mat_set_prec

ggml-ci

* Update ggml-cuda.cu

Co-authored-by: slaren <slarengh@gmail.com>

* Update ggml-cuda.cu

Co-authored-by: slaren <slarengh@gmail.com>

* cuda : ggml_cuda_op_mul_mat_cublas support F32 precision

* cuda : remove oboslete comment

---------

Co-authored-by: Ebey Abraham <ebeyabraham@microsoft.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
2023-12-18 19:27:47 +02:00
slaren
6744dbe924
ggml : use ggml_row_size where possible (#4472)
* ggml : use ggml_row_size where possible

ggml-ci

* ggml : move ggml_nbytes_split to ggml-cuda.cu
2023-12-14 20:05:21 +01:00
Georgi Gerganov
4d98d9a656
sync : ggml (SD ops, tests, kernels) (#4444)
* sync : ggml (SD ops, tests, kernels)

ggml-ci

* cuda : restore im2col

ggml-ci

* metal : fix accuracy of dequantization kernels

ggml-ci

* cuda : restore correct im2col

ggml-ci

* metal : try to fix moe test by reducing expert size

ggml-ci

* cuda : fix bin bcast when src1 and dst have different types

ggml-ci

---------

Co-authored-by: slaren <slarengh@gmail.com>
2023-12-13 21:54:54 +02:00
slaren
799a1cb13b
llama : add Mixtral support (#4406)
* convert : support Mixtral as LLAMA arch

* convert : fix n_ff typo

* llama : model loading

* ggml : sync latest ggml_mul_mat_id

* llama : update graph to support MoE

* llama : fix cur -> cur_expert

* llama : first working version

* llama : fix expert weighting in the FFN

* ggml : ggml_get_rows support 2D indexing [n_tokens, n_experts] (cpu only)

* ggml : add n_as argument to ggml_mul_mat_id

* ggml : fix ggml_get_rows to take into account ne02 / ne11

* metal : add more general support for ggml_get_rows + tests

* llama : add basic support for offloading moe with CUDA

* metal : add/mul/div use general kernel when src1 not cont

* metal : reduce the kernel launches for ggml_mul_mat_id

* ggml : get_rows : support non-contiguos tensors with gaps, generalize up to 3D

* ggml : update get_rows f16 and q

* cuda : support non-contiguous src1 in get_rows

* llama : offload missing ffn_moe_silu

* metal : fix ggml_get_rows to work with non-cont src1

* metal : add indirect mat-vec kernels for all quantization types

* llama : do not quantize expert gating tensors

* llama : add n_expert and n_expert_used to hparams + change quants

* test-backend-ops : add moe test

* cuda : fix get_rows when ncols is odd

* convert : determine n_ctx correctly

* metal : fix ggml_mul_mat_id for F32

* test-backend-ops : make experts more evenly probable (test_moe)

* test-backend-ops : cleanup, add moe test for batches

* test-backend-ops : add cpy from f32 -> all types test

* test-backend-ops : fix dequantize block offset

* llama : fix hard-coded number of experts

* test-backend-ops : simplify and disable slow tests to avoid CI timeout

* test-backend-ops : disable MOE test with thread sanitizer

* cuda : fix mul_mat_id with multi gpu

* convert : use 1e6 rope_freq_base for mixtral

* convert : fix style

* convert : support safetensors format

* gguf-py : bump version

* metal : add cpy f16 -> f32 kernel

* metal : fix binary ops for ne10 % 4 != 0

* test-backend-ops : add one more sum_rows test

* ggml : do not use BLAS with ggml_mul_mat_id

* convert-hf : support for mixtral-instruct (#4428)

* convert : typo fix, add additional hyperparameters, use LLaMA arch for Mixtral-instruct

* convert : use sentencepiece tokenizer for Mixtral-instruct

* convert : make flake8 happy

* metal : fix soft_max kernels

ref: 1914017863

* metal : limit kernels to not use more than the allowed threads

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Radek Pilar <github@mrkva.eu>
2023-12-13 14:04:25 +02:00
Richard Kiss
9494d7c477
english : use typos to fix comments and logs (#4354) 2023-12-12 11:53:36 +02:00
Georgi Gerganov
fe680e3d10
sync : ggml (new ops, tests, backend, etc.) (#4359)
* sync : ggml (part 1)

* sync : ggml (part 2, CUDA)

* sync : ggml (part 3, Metal)

* ggml : build fixes

ggml-ci

* cuda : restore lost changes

* cuda : restore lost changes (StableLM rope)

* cmake : enable separable compilation for CUDA

ggml-ci

* ggml-cuda : remove device side dequantize

* Revert "cmake : enable separable compilation for CUDA"

This reverts commit 09e35d04b1.

* cuda : remove assert for rope

* tests : add test-backend-ops

* ggml : fix bug in ggml_concat

* ggml : restore `ggml_get_n_tasks()` logic in `ggml_graph_plan()`

* ci : try to fix macOS

* ggml-backend : remove backend self-registration

* ci : disable Metal for macOS cmake build

ggml-ci

* metal : fix "supports family" call

* metal : fix assert

* metal : print resource path

ggml-ci

---------

Co-authored-by: slaren <slarengh@gmail.com>
2023-12-07 22:26:54 +02:00
Galunid
f23c0359a3
ci : add flake8 to github actions (python linting) (#4129)
Disabled rules:

* E203 Whitespace before ':' - disabled because we often use 'C' Style where values are aligned

* E211 Whitespace before '(' (E211) - disabled because we often use 'C' Style where values are aligned

* E221 Multiple spaces before operator - disabled because we often use 'C' Style where values are aligned

* E225 Missing whitespace around operator - disabled because it's broken so often it seems like a standard

* E231 Missing whitespace after ',', ';', or ':' - disabled because we often use 'C' Style where values are aligned

* E241 Multiple spaces after ',' - disabled because we often use 'C' Style where values are aligned

* E251 Unexpected spaces around keyword / parameter equals - disabled because it's broken so often it seems like a standard

* E261 At least two spaces before inline comment - disabled because it's broken so often it seems like a standard

* E266 Too many leading '#' for block comment - sometimes used as "section" separator

* E501 Line too long - disabled because it's broken so often it seems like a standard

* E701 Multiple statements on one line (colon) - broken only in convert.py when defining abstract methods (we can use# noqa instead)

* E704 Multiple statements on one line - broken only in convert.py when defining abstract methods (we can use# noqa instead)
2023-11-20 11:35:47 +01:00
Jiří Podivín
f7d5e97542
py : remove superfluous import statements (#4076)
Signed-off-by: Jiri Podivin <jpodivin@gmail.com>
Co-authored-by: Jiri Podivin <jpodivin@redhat.com>
2023-11-17 17:20:53 +02:00
Galunid
36eed0c42c
stablelm : StableLM support (#3586)
* Add support for stablelm-3b-4e1t
* Supports GPU offloading of (n-1) layers
2023-11-14 11:17:12 +01:00
Georgi Gerganov
4760e7cc0b
sync : ggml (backend v2) (#3912)
* sync : ggml (backend v2) (wip)

* sync : migrate examples and llama.cpp to dynamic graphs (wip)

* sync : update tests + fix max op params to 64

ggml-ci

* sync : ggml-cuda

ggml-ci

* llama : fix save/load state context size

ggml-ci

* sync : try to fix build on tvOS

* sync : pass custom graph sizes in training examples

* sync : update graph copies to new ggml API

* sync : update sync-ggml.sh with new files

* scripts : fix header in sync script

* train : fix context size calculations

* llama : increase inference graph size up to 4096 nodes

* train : allocate grads for backward graphs

* train : allocate grads for gb_tmp
2023-11-13 14:16:23 +02:00
Georgi Gerganov
207b51900e
ggml : move FP16 <-> FP32 code to ggml-impl.h (#3861)
* ggml : move FP16 <-> FP32 stuff to ggml-impl.h

ggml-ci

* tests : fix ARM build

* ggml : explicitly initialize deprecated type traits

* ggml : add math.h to ggml-impl.h

* ggml : remove duplicate static assert macros

* ggml : prefix lookup tables with ggml_

ggml-ci

* ggml-impl : move extern "C" to start of file
2023-10-30 19:19:15 +02:00
Galunid
daab3d7f45
Add more tokenizer tests (#3742)
* Add more tokenizer tests

* Add starcoder

* Update test vocab files

* Restrict bpe tokenizer tests to unicode planes

* Update comment

* Comment cosmetics

* Remove bloom vocab/test
2023-10-24 09:17:17 +02:00
goerch
9e70cc0322
Add test for MPT tokenization (#3728)
* Add test for MPT tokenization

* Revert code motion

* Remove unnecessary restriction in test case

* Clarify logic in conversion
2023-10-22 21:21:42 +02:00
Georgi Gerganov
d1031cf49c
sampling : refactor init to use llama_sampling_params (#3696)
* sampling : refactor init to use llama_sampling_params

* llama : combine repetition, frequency and presence penalties in 1 call

* examples : remove embd-input and gptneox-wip

* sampling : rename penalty params + reduce size of "prev" vector

* sampling : add llama_sampling_print helper

* sampling : hide prev behind API and apply #3661

ggml-ci
2023-10-20 21:07:23 +03:00
Qin Yue Chen
8cf19d60dc
gguf : support big endian platform (#3552)
* check whether platform is 390x if yes->do not import immintrin.h

* support s390x big endian

* support --bigendian option for s390x
1. verified with baichuan7b-chat with float 16 on s390x
2. verified with baichuan7b-chat
3. verified with chinese-alpaca-2-13b-f16

* update format based on editor-config checker result

* Update convert-baichuan-hf-to-gguf.py

* 1. check in ggml.c if endianess is not match
2. update GGUF version
3. change get_pack_prefix to property
4. update information log

* always use "GGUF" as beginng of GGUF file

* Compare "GGUF" with file header char by char
1.  Set GGUF_MAGIC to "GGUF" string instead of int value
2. Compare "GGUF" char by char to ensure its byte order
3. Move bytes swap code from convert.py to gguf.py write_tensor_data

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-10-20 14:19:40 +03:00
goerch
233fc1c69f
Minor improvements in GPT2 tokenizer (#3567)
* Fixing minor bugs in bpe_gpt2_preprocess

* Don't add bos token in test
2023-10-10 18:59:52 +02:00
Georgi Gerganov
f93af02488
sync : ggml (conv 1d + 2d updates, UB fixes) (#3468)
* sync : ggml (conv 1d + 2d updates)

ggml-ci

* ggml : fix UB in q5_0 and q5_1 quantize code

ggml.c:1033:39: runtime error: left shift of 1 by 31 places cannot be represented in type 'int'
SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior

ggml.c:1081:39: runtime error: left shift of 1 by 31 places cannot be represented in type 'int'
SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior

ggml-ci

* tests : fix UB in test-quantize-perf
2023-10-04 15:29:58 +03:00
goerch
ff5a3f0c09
Work on the BPE tokenizer (#3252)
* Work on the BPE tokenizer

Tokenizer tests work for Falcon-7B

* Try to fix build problem

* Fix debug assertion failure

* Fix MSVC Unicode BOM problem

* Cleanup and an improvement

* Fix compiler warning

* Cleanup

* Test doesn't work over the full range of Unicodes

* Update .gitignore and Makefile

* Another Makefile rule

* Testing Aquila

* Moving byte decoding back to `token_to_piece` ...

... because everyone is using it.

* Guarding some unusable code pathes

* Streamlining code and adding some more assertions

Important change: I'm classifying added tokens as control tokens now for BPE.

* Adding a comment

* Adding another assertion

* Fixed vocabulary guarding assertions

* Fix PR for recent change

* Fix PR for recent change

* Fix for compiler warning

* Fix PR for recent change

* Fix PR for recent change

* Fix PR for recent change

* Fix for compiler warning

* Fixes for more compiler warnings

* Remove unused code

* Fix initialization of static maps

* Add scores and token types back, adapt gptneox

* Update llama.cpp

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update unicode.h

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Update unicode.h

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* Ported Starcoder and added some assertions

* Fix coding style

* Apply @jploski 's fix for missing tokens

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-10-03 09:16:26 +02:00
Cebtenzzre
bc39553c90
build : enable more non-default compiler warnings (#3200) 2023-09-28 17:41:44 -04:00
slaren
16bc66d947
llama.cpp : split llama_context_params into model and context params (#3301)
* llama.cpp : split llama_context_params into model and context params

ggml-ci

* fix metal build

* fix freq_base/scale default to model value

* llama-bench : keep the same model between tests when possible

* move n_threads to llama_context_params, add n_threads_batch

* fix mpi build

* remove kv_size(), cuda scratch fixes

* remove low-vram option

* add n_threads_batch to system info, refactor to get_system_info()

* add documentation about --threads-batch to the READMEs

* llama-bench fix

* main : fix rope freq/scale warning

* llama.cpp : add llama_get_model
common : add llama_tokenize from model

* remove duplicated ctx/model functions

ggml-ci

* cuda : print total VRAM used
2023-09-28 22:42:38 +03:00
xaedes
0e76a8992c
train : finetune LORA (#2632)
* fix track_max_mem in forward_batch_wo_cache_flash_attn_train

* remove unnecessary Adam(W) optimizer tensors.

reduces optimizer memory overhead from 7*modelsize to 2*modelsize.

additionally allows to optimize models with more than 2^31 parameters by replacing int with int64_t.

bumps training checkpoint file version, but old checkpoints can still be read.
new version with less tensors is saved.

* add gradient clipping to AdamW

* Fix reset of unused g->nodes and g->grads to NULL

* implement gradient checkpointing for training

reduces memory overhead from O(n_layer) to O(sqrt(n_layer))

as explained in readme of https://github.com/cybertronai/gradient-checkpointing

* remove unused compute buffer 3

* add and use function ggml_build_backward_expand to avoid stack overflows with large maximum number of nodes

GGML_API void ggml_build_backward_expand(struct ggml_context * ctx, struct ggml_cgraph * gf, struct ggml_cgraph * gb, bool keep);

* change AdamW decay parameter to work like the torch AdamW decay parameter

It is now relative to Adam learning rate `alpha*sched`.
Before that it was relative to `sched` only.

`alpha` being the maximum learning rate and `sched` being a scaling parameter in [0..1]

* change default AdamW weight decay parameter used in training to 0.1 as used in nanoGPT

* change default AdamW weight decay parameter defined in ggml to 0.0, making Adam default instead of AdamW

btw: the default weight decay parameter for torch.optim.AdamW is 0.01

* bug fixes for cross entropy loss

ggml_cross_entropy_loss: sums where not correctly added in workload of each thread
ggml_cross_entropy_loss_back: simplify backward process, reducing numerical issues

guard usage of exp f16 lookup in cross entropy by #define GGML_CROSS_ENTROPY_EXP_FP16

cross entropy loss is only used once during training, but it is quite sensitive to numerical errors introduced by exp-f16-lookup.
so exp-f16-lookup for cross entropy loss is disabled by default, trading better gradients for very slightly worse runtime performance.

* fix test-grad0 for cross_entropy_loss

the second argument to cross_entropy_loss must sum up to 1 for each row

* fix test-grad0 for soft_max

dont use only sum as aggregation, because sum of softmax is always 1 -> finite differences should not work
instead use sum(log(soft_max()*(1-eps)+eps)); use eps to avoid log(0)

* improve finite differences of test-grad0 by using double instead of float

* change cross_entropy_loss to output average over all rows

this helps keeping the loss and gradients in a sane range

* improve gradient checkpointing

sqrt(n_layers) is only the best checkpoint step when mem size of checkpoints and mem size of layers are equal.
since layers require more memory than the single-tensor-checkpoint we use, the optimal values are compute different:

```
  given: n, u, v
  objective: minimize(a*u+b*v) where a*b=n, a>0, b>0
  b=n/a
  minimize(a*u+v*n/a)
  diff(a*u+v*n/a, a) = u - (v*n/a)/a
  diff(a*u+v*n/a, a) == 0
  u - (v*n/a)/a == 0
  u == v*n/(a*a)
  u*a*a = v*n
  a*a = v*n/u
  a = sqrt(n*v/u)
```

this change results in more checkpoints, requiring less layers to store between checkpoints, overall improving memory usage.

* disable gradient checkpointing debug output

* llama : fix rope usage in train-text-from-scratch after ChatGLM change

* add more training parameters:

--enable-restart N         Only for Adam optimizer. Enable restarts of cos-decay
--disable-restart N        Only for Adam optimizer. Disable restarts of cos-decay
--opt-past N               Number of optimization iterations to track for delta convergence test. Disabled when zero.
--opt-delta N              Maximum delta for delta convergence test. Disabled when <= zero.
--opt-max-no-improvement N Maximum number of optimization iterations with no improvement. Disabled when <= zero.
--adam-epsf N              AdamW epsilon for convergence test. Disabled when <= zero.
--adam-min-alpha N         Adam minimum learning rate alpha, usually 0.1 * alpha

* replace memcpy with reshape operation so that the graph is not cut at the input

this makes it possible to store other values into the input tensor and then simply recompute the graph without rebuilding it

* remove unused function argument from get_example_targets_batch

* measure and print total training time

* add optimization callback to ggml_opt_resume_g

this callback is called before each iteration with custom data and pointer to learning schedule parameter (only used in Adam(W)).

can be used for dynamic learning schedule and setting input data for batches before each iteration

* use optimization callback in training

allows dynamic learning schedule and different batch data for each iteration without relying on low n_iter and high n_examples parameters

reduces runtime by avoiding restart of optimization function and improves training convergence by providing a different batch for each iteration

* add minimum number of tensor dimensions to apply weight decay (default 2)

this allows to not apply weight decay to bias parameters

* rename training parameter cos-decay-alpha to cos-decay-min and clarify that adam-min-alpha also applies to warmup

* fix increase of model.train_samples and model.train_tokens

now that each optimizer iteration gets its own batch we need to multiply by number of opt iterations

* change sampling parameters for prediction after training to defaults of common.h

and clarify what is context for prediction and what are generated tokens

* tighten abs error bounds for cross_entropy_loss in test-grad0

* add conditional compilation of using F16 exp in flash attention

uncomment `// #define GGML_FLASH_ATTN_EXP_FP16` to enable usage of f16 exp in flash attention

* tighten abs error bounds for flash_attn in test-grad0

* tighten abs error bounds for sqrt in test-grad0

* remove out-commented vectorized code of opt_adam

the vectorized code might be bit faster for low number of parameters, but it had a big memory usage overhead

* ggml : update ggml_rms_norm_back with configurable eps

* llama training : fix ggml_rms_norm_back calls to pass configurable eps

* remove trailing whitespace

* add train function using automatic gradient checkpointing backward pass and allocator

* in train function replace add_inplace by regular add

because using add_inplace seems to result in different gradients

* don't use allocate hash_map on context

because the context has no_alloc=True when using memory allocator resulting in NULL data pointers

* correctly clone reshape and permute operations by also cloning tensor->nb values

* fix variable name and add missing type cast

* terminate recursive tensor cloning when reaching tensor without src tensors

* correctly clone view tensors by setting data pointers

without this the checkpointing would only work when being used together with memory allocator

* fix variable names

* swap arguments to commutative ops to be the same as in `forward_batch_wo_cache_flash_attn`

* add input tensors as checkpoints

so that recursive tensor cloning of gradient checkpointing terminates on input tensors

* fix variable name and add missing boolean negation

* make sure some tensors are not reallocated by inserting new temporary nodes depending on them:

output and parameter gradient tensors need to be available at the end of the graph execution

parameter gradient tensors also need to be available before the graph execution because they are set to zero before each optimizer iteration

checkpoint tensors are allocated all together to reduce memory allocator fragmentation

afterwards, in addition to the temporary nodes, we also need to reset the temporary leafs

* fix ASSERT to work with zero layers

* add training options whether to use allocator and/or unified training function

* integrate unified training function which may use memory allocator

the unified training function also supports arguments whether to use flash attention and/or gradient checkpointing

* format name of cloned tensors with " (clone)" suffix

* set names for tensors in unified train function for easier debugging

* allocate graph on context using ggml_new_graph

* remove handwritten training functions

* remove unused training parameters "use_scratch" and "use_unified"

* remove trailing whitespace

* remove unused train params: mem_compute1_gb & mem_compute2_gb

mem_compute_gb is used for compute when automatic memory allocator is not enabled, otherwise it can be very small to only hold the tensor definitions
mem_compute0_gb is used for automatic memory allocator (as long as measurement of max required size is not implemented)

* remove unused forward_batch function

* add debug asserts in ggml_allocr_alloc to some common pitfalls when using this function directly

* only use ggml_allocr_alloc when tensor has NULL data and is no view

* fix test when to create temporary backward graph

temporary backward graph is only necessary when using checkpointing

* fix memory "leak" in optimizers

each iteration a new cplan with new memory for work data was allocated.
now cplan creation only happens at the start of optimization, with each iteration reusing the cplan and its work data.

* reverse order of for loop in ggml_build_backward_expand to save memory when using gradient checkpointing and allocator

with this loop order gradient checkpointing with allocator on 16 layer model saves 13% memory; 2 layer memory it saves 2% memory.

the computation results are the same

* add API functions to access llama model tensors

* add stub example for finetuning, based on train-text-from-scratch

* move and remove code

* add API functions to access remaining model parameters:

mult, head and rot

* first draft for LORA finetune training

* remove const model and layer arguments in API functions for accessing model tensors

* bug fixes to make finetune compile

automatic allocator does not work yet

* add debug prints for training memory improvements

* fix names of lora tensors

* avoid stack overflow resulting from big ggml_cgraph

replace stack allocation and ggml_build_forward by ggml_new_graph in combination with ggml_build_forward_expand

* replace llama API functions to get model tensors by one function to get model tensor by name

LLAMA_API struct ggml_tensor * llama_get_model_tensor(struct llama_model * model, const char * name);

* remove unused call to not existing llama_get_layer_from_model

* implement ggml_compute_forward_out_prod_q_f32

* remove trailing whitespace

* add lora finetune support on quantized base model tensors

* add ggml_add_cast API function

this function works like ggml_add, but accepts a data type for the resulting tensor.
only supported for quantized src0 input.

* use ggml_add_cast in finetuning

lora-applied weights will now have data type F32, which improves gradients when finetuning quantized base models

* bug fix: actually use result type passed to ggml_add_cast

* make sure base model tensors data cannot be used in viewable operations

memory allocator would try to make lora application inplace on base model tensors.
since those are memory mapped this will result in memory access violations

* fix bug in ggml_out_prod which resulted in wrong n_dims of result tensors

* avoid keeping in memory ALL of the gradients

The problem here stems from ggml_graph_reset. This function is called in the optimization function, before each graph computation, to reset the gradients to zero. This required a unique memory slot for each gradient: allocating memory from a previosly freed memory location might lead to non-zero input gradients.

During ggml_compute_backward the gradients are build stepwise by adding or substracting new values, starting from a OP_NONE tensor which needs to contain zero-values. This requires the graph reset.

To avoid this I now remember in ggml_build_backward_expand the original OP_NONE gradient tensors in a hash table, which is passed to ggml_compute_backward. There instead of using add (or sub or similar) I test whether the existing gradient to be changed is a zero-valued-tensor by looking up its existence in the hash table. When it is such a zero-tensor it will not be modified, but replaced by the value to be added, otherwise the regular add (not inplace, allocator will take care of this) will be used. This way none of those zero-tensor values will be necessary in the final backward graph and more importantly they won't need a unique memory slot, just to make them zero.

* remove trailing whitespace

* remove debug prints and function to compute tensor data hash

* improve optimization iteration prints

* adjust maximal values to support finetuning 3B models

* change default finetune params lora_r and lora_alpha to match the n_rank parameters of 4

* bug fix: make sure finetune input gradient is allocated at begin and kept until end

* remove unnecessary src tensor from ggml_get_rows_back

we don't need data of src[2] for computation, only to setup the correct output shape.
remove dependency on src[2], so that allocator can work more freely.

the computational graph is still completely determined, because the output shape is naturally included.
this is similar to how ggml_reshape does it.

* remove unnecessary src tensor from ggml_repeat & ggml_repeat_back

we don't need data of src[1] for computation, only to setup the correct output shape.
remove dependency on src[1], so that allocator can work more freely.

the computational graph is still completely determined, because the output shape is naturally included

* resolve todo

allocator will only make it inplace when they are of the same type

* mixing multiple LORA adapters is now possible

pass more than one '--lora FNAME' argument to apply more than one LORA.
use '--lora-scaled FNAME S' when you want to specify a user-defined scale for an adapter.

* add option to save finetune output every N iterations

* also save latest finetune output with ITERATION="LATEST" and print where files are saved

saving with LATEST makes it easier to resume training from the latest checkpoint
the string "LATEST" can be configured with command line option "--fn-latest STR"

* update checkpoint train stats before saving via "--save-every"

* add command line option `--rank-wo N` for rank of wo tensor

* update finetune README

* fix dump_non_result_info_yaml to output multiple lora adapters

* bug fix: replace GGML_TYPE_SIZE[t] by ggml_type_size(t)

* replace llama_n_mult by llama_n_ff

* finetune bug fixes to compile with merged in code from master

* remove prediction related code to reduce duplicated code with main

use main instead

* reduce large memory overhead in train-text-from-scratch

all gradients had to be pinned so that graph_reset works correctly.
this is no longer necessary with the changes to ggml_compute_backward introduced in this PR.

* add comment explaining why finetune checkpoints are allocated in one block

* make default value of float member a float literal

* handle rms_norm and rope parameters the same as in train-text-from-scratch

* remove unused code

* remove vocab related code as it is unnecessary

* add LLM_KV_TRAINING_TYPE to train-text-from-scratch checkpoints

so that they can be differentiated from lora finetune checkpoints

* add gguf constants and load/save functions from train-text-from-scratch

* add load & save lora finetune checkpoints via gguf

* add python script to convert old finetune checkpoint files to gguf

* remove old checkpoint save & load code

* remove code to print data checksums which was used to verify correctness of new gguf code

* omit tokenization when training is disabled, only save llama lora adapter

training can be disabled by passing '-n 0' to finetune

* remove trailing whitespace

* update README.md

* implement ggml_compute_forward_repeat_f16

* avoid stack overflow of large cgraphs in test-grad0

* add ggml API functions ggml_unravel_index, ggml_get_i32_nd and its analogs for set and for f32

ggml_get_i32_1d, ggml_set_i32_1d, ggml_get_f32_1d, ggml_set_f32_1d now support non-contiguous tensors.
in case of non-contiguous tensor, the 1d index is unraveled into a multi index using ggml_unravel_index to be passed to '_nd' function equivalent.

this fixes a bug in test-grad0 which happens due to ggml_build_backward not building purely contiguous tensors anymore

* increase test-grad0 context mem size to accommodate for bigger cgraph

* add sanity check to ggml_compute_backward, asserting the correct shape of gradients

* fix ggml_acc_or_set to return tensor of correct shape

* remove unused 'inplace' argument from ggml_compute_backward function

inplace operations to add gradients are no longer created by ggml_compute_backward
use allocator to automatically make inplace operations

* add missing argument 'int i0' to ggml_get_i32_nd & ggml_set_i32_nd header declarations

* fix error message in ggml_allocr_alloc to display actual max_avail

* fix check_gradient

ggml_build_backward_expand was previously replaced by ggml_build_backward, but the assignment of forward graph to backward graph missing

* use tensor->view_src instead of ggml_is_view and get_view_source

* move gradient checkpointing code into ggml, new API function:

// build gradient checkpointing backward graph gb for gf using provided checkpoints
// gb_tmp will contain original backward graph with rewritten backward process nodes,
// but without the second forward pass nodes.
GGML_API void ggml_build_backward_gradient_checkpointing(
        struct ggml_context   * ctx,
        struct ggml_cgraph    * gf,
        struct ggml_cgraph    * gb,
        struct ggml_cgraph    * gb_tmp,
        struct ggml_tensor  * * checkpoints,
        int                     n_checkpoints);

* replace custom data getters and setters by ggml functions

* train-text-from-scratch can train (full finetune) gguf models

just pass the gguf model via `--checkpoint-in FN`.
after this, to continue training, pass the generated checkpoint instead of the original gguf model.

tested with smaller models, bigger models may exceed available memory.
use (LORA) finetune for those.

* remove trailing whitespace

* add option to save train-text-from-scratch output every N iterations

* update README.md

* fix warnings

* fix warnings

* remove finetune option to disable allocator

the allocator should always be used.
by making sure that it is always used it gets easier to implement automatic memory requirements computation

* add tensor checkpoints only when gradient checkpointing is enabled

* initialize opt ggml context if none was provided

* add ggml-alloc API function 'ggml_allocr_max_size' to get max size of alloc

GGML_API size_t ggml_allocr_max_size(struct ggml_allocr * alloc);

* finetune: automatically allocate all memory and changes to command line options

remove '--n_examples N' parameter, as it no longer makes sense to call optimization process multiple times in a loop.
add '--only_write_lora' command line option: will skip tokenization and training, to only write a llama.cpp comptabile LORA adapter.
remove memory buffer related command line options.
improve iteration console output.

* add finetune to Makefile

* update README.md

* print time per iteration and estimate remaining time

* increase measured alloc size by tensor_alignment

ggml_allocr_reset will reduce the given size by up to tensor_alignment-1

* fix README.md

* add some more allocator debug prints

* bug fix, probably solves the 'ggml_allocr_alloc: not enough space in the buffer' issue

* revert last commit

"bug fix, probably solves the 'ggml_allocr_alloc: not enough space in the buffer' issue"

"alloc was freeing an externally allocated tensor, because it calculated the end of allocator memory as alloc->data + alloc->max_size instead of alloc->data + alloc->size."

This is intentional to reduce the risk of freeing external tensors when measuring. Unless max_size is not properly calculated, I don't see why this is an issue.

* remove unnecessary "0x" before "%p" output

* move measurement memory segment to upper region of the address space

* update README.md

* fix printf format warnings

* add missing gguf_free in load_checkpoint_lora_file

* load default rms_norm and rope parameters from base model

* add gradient accumulation

specify number accumulation steps with '--grad-acc N'.
this will simulate a bigger batch size of grad_acc*batch.

* fix tracking of train_samples and train_tokens

* build : fix compile warnings

* ggml : fix L-BFGS linesearch loop

* improve finetune time measurement

fix printf warnings on system where int64_t is (long int).
change time datatypes to double because values get big with long training times.
exclude file saving from time measurement.
converge faster to actual time per iteration by removing very small first duration before first iteration was performed.
fix bug in output of total training time, the reported value was 1000 times to small.

* specify default lora rank with '--lora-r N'

'--lora-r N' will specify default rank for all tensors
'--rank-wq N', etc. will override this default rank for specific tensor types.

* fix gradient accumulation bug where the same batch was used for each microstep

* fix gradient accumulation bug where the same batch was used for each microstep

* support grouped-query-attention in ggml_flash_attn and ggml_flash_attn_back

k and v can now be repeated in q along ne[2]

in forward pass just use modulo to compute k and v indices, like ik2 = iq2 % nek2.

in backard pass this won't work as easy, because multiple threads will compete to accumulate to the same k->grad[:,ik1,ik2,ik3] and v->grad[:,iv1,iv2,iv3].
so we change the parallelization over q rows to be over k rows. this ensures non-overlapping (ik2,ik3) across threads.
in each thread we then iterate over the number of repetitions of k/v in q to compute iq2 as iq2 = ik2 + irep*nek2.

since ne2 is not the same for q,k and v we also change how the gradients are concatenated into the result tensor.
additionally the offsets of gradq, gradk and gradv in the result tensor are now memory aligned.

we also simplify the compute_backward part of flash_attn to use ggml_reshape instead of switching over the number of dimensions.
this needs a small change to ggml_reshape, removing the assertion of second argument to be contiguous.
since only the shape (ne) of the second reshape argument is of relevance, its memory layout (nb) is irrelevant -> it can very well be non-contiguous.

change test-grad0 to also test for repeated k/v in q.

this changes the rng and now results in small gradient differences in softmax. these solely come from using f16 exp table lookup in forward softmax: when temporarily changing softmax to use actual exp function, the reported gradient differences go away. gradient differences coming solely from f16 table lookup are acceptable.
added a note to explain this.

* add llama API functions to get grouped-query-attention n_head parameter 'n_head_kv'.

* fix finetune to support grouped-query-attention (using flash-attention)

note: ggml changes to ggml_out_prod are necessary to support grouped-query-attention without flash-attention.

* support broadcastable a in out_prod(a, b) and backward pass of broadcasting mul_mat(a, b)

* test broadcasting mul_mat backward pass

* decouple random number generator of each operation test

when changing one test the rng of others tests is not influenced anymore

* add comment briefly describing what ggml_repeat_back does

* simplify broadcasting mul_mat backward using ggml_repeat_back

* add cgraph evaluation order member and corresponding enum type

this controls in which order ggml_build_forward visits source nodes.
by default the nodes are visited left to right, i.e. src[0] first.
in some cases it is beneficial for ggml-alloc to visit in a different order.
two possible orders are supported: left-to-right (src[0] first) and right-to-left (src[0] last).

* measure max compute size for each cgraph eval order and use best order

this can bring huge memory savings:
e.g. codellama-34b with n_ctx=64, n_batch=1 goes from 92927.8mb down to 4627.6 MB

* remove unused command line options

* add sample start patterns and options to force new or by default resume last shuffling

* update shuffle rng state on reshuffle

* exclude known zero values from computations in flash_attn_f32 & flash_attn_back_f32

* remove probably unnecessary exception type flags from stringstream

* pass correct max number of tokens to llama_tokenize

* account for possible leading whitespace that will be added by tokenizer
e.g. '\t' will be tokenized by llama spm tokenizer to [29871, 12]

* use unrolled vec_mad in out_prod

y is vec_mad result vec.
x is vec_mad input vec.
v is vec_mad input scalar.

ggml_vec_mad_f32_unroll will internally loop over x and v with same y.

GGML_VEC_MAD_UNROLL is by default defined to 32.

This value is empirical optimized using performance test runs of out-prod in openllama-3b finetune with 256 context length and batch size 1. It gives 23% performance boost for out_prod.

Full measurements of out-prod runtime in ms:
	unroll_xv	unroll_yv
1	67014.643	87826.469
2	77117.552	89077.656
4	72091.311	109121.657
8	61077.543	88678.334
16	56914.67	79514.947
24	59024.595	84350.254
28	55952.446	83368.73
32	51476.658	85177.745
36	55973.792	84659.92
40	55139.616	93844.738
48	60736.392	93330.267
64	99856.878	116994.99

Second column is when unrollying yv instead of xv

* set lora_alpha to value of lora_r if it is not set via command line

otherwise only changing lora_r will change scaling of lora adapter used in prediction

* reshuffle original sample order instead of the previous shuffled order

otherwise resumed reshuffle will not result in same sample order

* block tiling for out-prod inspired by mul-mat

block sizes are empirically optimized

roughly doubles the flops of out-prod

* exclude some more known zero values from computations in flash_attn_f32 & flash_attn_back_f32

* add static keywords

* remove outcommented old code

* update train-text-from-scratch with tokenization, sample selection and shuffling from finetune

* remove lbfgs related train parameters

* move common train functions into common/train.[h|cpp]

* move train state into struct train_state

* move train data saving code into callback to unify code of opt_callback

train_params are still different in finetune and train-text-from-scratch, so it can't yet be moved to train.h|cpp

* move common train params into common/train

* move common opt_callback into common/train

* fix consume_common_train_arg

* save and load head_count_kv in lora checkpoints

* increase train_samples by used_samples instead of number of batches

on batch can contain more than one sample when option "fill_with_next_samples" is used

* fix usage of llama_tokenize

* remove static from process_escape since we need it exposed in header

* fix code formating of long function declarations

* fix condition in load_train_state_gguf

* use die("msg") instead of replace GGML_ASSERT(!"msg") or throw std::runtime_error("msg")

* fix saving and loading of training type

* remove terminating '\0' from tokenization

(llama_tokenize is now passed the string length instead of relying on terminating '\0')

* fix compile warnings

* fix compile warnings

* use new/delete for train_state instead of malloc/free

using malloc may result in seg faults when trying to assign string fields

* assert that sample_count > 0, avoiding division by zero

* fix frand to return value in interval [0,1)

* add train option "--sample-random-offsets"

Use samples beginning at random offsets.
The offset is only applied to the first sample in each batch context window.
Together with "--fill-with-next-samples" this may help for training endless text generation.

For example given a dataset containing samples "abcd", "ABCD", "0123".
With context size of 8 and options "--fill-with-next-samples", "--no-separate-with-eos", "--no-separate-with-bos",
the context windows of batches could only be filled with "abcdABCD", "ABCDabcd", "0123abcd", etc.

With "--sample-random-offsets" it can also be filled with "23abcdAB", "bcd0123A", etc.

* deduplicate code into function

* remove n_rot hparam, as it must always be hparam.n_embd_head()

* align code

* assert correct base model tensor shapes

* move some params from lora hparams into model hparams and load model params from gguf

this equalizes the model definition in finetune and text-from-scratch and removes the need for additional llama api functions to get model parameters

* remove now unnecessary llama API functions to get model params that where added by this PR

* train-text-from-scratch: automatically allocate model tensors, remove option '--mem-model N'

* train-text-from-scratch: automatically allocate opt context

* train-text-from-scratch: automatically allocate input tensors

* train-text-from-scratch: automatically allocate compute memory

* remove unused options and equalize train-text-from-scratch with finetune

* initialize opt->loss_after with zero

* add export-lora program

* remove trailing whitespace

* add export-lora build in Makefile

* remove unused struct tensor_info from export-lora

* add export-lora build dependency to llama

because it depends on common, which depends on llama

* update finetune README.md

* cancel optimization when specified number of epochs is completed

* improve handling of export-lora arguments

print errors and warnings when files could not be read or created

* Fix export-lora.cpp "not enough space in the context's memory pool" (#1)

* Fix export-lora.cpp "not enough space in the context's memory pool"

Without this patch, export-lora would sometimes error with "not enough space in the context's memory pool (needed 656784, available 656800)".

* increase required context size by 5*GGML_MEM_ALIGN instead of plain 16

---------

Co-authored-by: xaedes <xaedes@gmail.com>

* improve handling of not yet supported tensor types

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: meatbag-18a <145869052+meatbag-18a@users.noreply.github.com>
2023-09-28 21:40:11 +03:00
Georgi Gerganov
ec893798b7
llama : custom attention mask + parallel decoding + no context swaps (#3228)
* tests : verify that RoPE is "additive"

* llama : replace ggml_diag_mask_inf with ggml_add (custom -inf mask)

* ggml : ggml_rope now takes a vector with positions instead of n_past

* metal : add rope_f16 kernel + optimize cpy kernels

* llama : unified KV cache + batch inference API

* llama : add new llama_decode() API that works with llama_batch

* llama : add cell_max heuristic for more efficient kv_cache

* llama : extend llama_kv_cache API

* llama : more robust cell_max heuristic + wip shift

* metal : disable concurrency optimization

* llama : add llama_kv_cache_shift_seq + no more context swaps

* llama : apply K-cache roping for Falcon and Baichuan

* speculative : fix KV cache management

* parallel : example for serving multiple users in parallel

* parallel : disable hot-plug to avoid cache fragmentation

* fixes : speculative KV cache + llama worst-case graph

* llama : extend batch API to select which logits to output

* llama : fix worst case graph build

* ggml-cuda : update rope implementation for parallel decoding (#3254)

* ggml-cuda : update rope implementation for parallel decoding

* better solution for p0 computation

* fix rope

* simpler rope implementation

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* make : add parallel to build + fix static functions in llama.cpp

* simple : fix token counting

* parallel : various improvements

* llama : fix cell_max logic + rename functions

* parallel : try smaller batches when the KV cache is fragmented

* parallel : fix sequence termination criteria

* llama : silence errors KV cache errors

* parallel : remove new line from prompt

* parallel : process system prompt once + configurable paramters + llama API

* parallel : remove question with short answers

* parallel : count cache misses

* parallel : print misses on each request

* parallel : minor

* llama : fix n_kv to never become 0

* parallel : rename hot-plug to continuous-batching

* llama : improve llama_batch API + simplify parallel example

* simple : add parallel decoding support

* simple : improve comments + free batch

* ggml-cuda : add rope f16, restore performance with parallel decoding (#3272)

* ggml-cuda : add rope f16, restore performance

* offload KQ_mask with all models

* fix rope shift

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* llama : disable MPI for now

ggml-ci

* train : make KQ_pos memory buffer permanent via dummy scale op

* ggml : revert change to ggml_cpy, add ggml_cont_Nd instead (#3275)

ggml-ci

* parallel : fix bug (extra BOS) + smaller token_prev array

* parallel : fix cases where the input prompts can overflow the batch

* parallel : add disabled experimental batch chunking in powers of two

* llama : llama.h formatting + comments

* simple : add README.md

* llama : fix kv cache heuristic when context is less than 32

* parallel : fix crash when `-n -1`

* llama : simplify returns if/else branches

* metal : use mm kernels for batch size > 2

* examples : utilize new llama_get_logits_ith()

* examples : add example for batched decoding

* examples : do not eval prompt 2 times (close #3348)

* server : clear the KV cache beyond n_past before llama_decode

* server : avoid context swaps by shifting the KV cache

---------

Co-authored-by: slaren <slarengh@gmail.com>
2023-09-28 19:04:36 +03:00