ayo/ipex-llm
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1
2096 commits 1 branch 0 tags 37 MiB
Commit graph

832 commits

Author SHA1 Message Date
Guancheng Fu
5857a38321 [vLLM] Add option to adjust KV_CACHE_ALLOC_BLOCK_LENGTH (#9782) 
* add option kv_cache_block

* change var name
 2023-12-28 14:41:47 +08:00
Ruonan Wang
99bddd3ab4 LLM: better FP16 support for Intel GPUs (#9791) 
* initial support

* fix

* fix style

* fix

* limi esimd usage condition

* refactor code

* fix style

* small fix

* meet code review

* small fix
 2023-12-28 13:30:13 +08:00
Yishuo Wang
7d9f6c6efc fix cpuinfo error (#9793) 2023-12-28 09:23:44 +08:00
Wang, Jian4
7ed9538b9f LLM: support gguf mpt (#9773) 
* add gguf mpt

* update
 2023-12-28 09:22:39 +08:00
Cengguang Zhang
d299f108d0 update falcon attention forward. (#9796) 2023-12-28 09:11:59 +08:00
Shaojun Liu
a5e5c3daec set warm_up: 3 num_trials: 50 for cpu stress test (#9799) 2023-12-28 08:55:43 +08:00
dingbaorong
f6bb4ab313 Arc stress test (#9795) 
* add arc stress test

* triger ci

* triger CI

* triger ci

* disable ci
 2023-12-27 21:02:41 +08:00
Kai Huang
40eaf76ae3 Add baichuan2-13b to Arc perf (#9794) 
* add baichuan2-13b

* fix indent

* revert
 2023-12-27 19:38:53 +08:00
Shaojun Liu
6c75c689ea bigdl-llm stress test for stable version (#9781) 
* 1k-512 2k-512 baseline

* add cpu stress test

* update yaml name

* update

* update

* clean up

* test

* update

* update

* update

* test

* update
 2023-12-27 15:40:53 +08:00
dingbaorong
5cfb4c4f5b Arc stable version performance regression test (#9785) 
* add arc stable version regression test

* empty gpu mem between different models

* triger ci

* comment spr test

* triger ci

* address kai's comments and disable ci

* merge fp8 and int4

* disable ci
 2023-12-27 11:01:56 +08:00
binbin Deng
40edb7b5d7 LLM: fix get environment variables setting (#9787) 2023-12-27 09:11:37 +08:00
Kai Huang
689889482c Reduce max_cache_pos to reduce Baichuan2-13B memory (#9694) 
* optimize baichuan2 memory

* fix

* style

* fp16 mask

* disable fp16

* fix style

* empty cache

* revert empty cache
 2023-12-26 19:51:25 +08:00
Jason Dai
361781bcd0 Update readme (#9788) 2023-12-26 19:46:11 +08:00
Yuwen Hu
c38e18f2ff [LLM] Migrate iGPU perf tests to new machine (#9784) 
* Move 1024 test just after 32-32 test; and enable all model for 1024-128

* Make sure python output encoding in utf-8 so that redirect to txt can always be success

* Upload results to ftp

* Small fix
 2023-12-26 19:15:57 +08:00
WeiguangHan
c05d7e1532 LLM: add star_corder_15.5b model (#9772) 
* LLM: add star_corder_15.5b model

* revert llm_performance_tests.yml
 2023-12-26 18:55:56 +08:00
Ziteng Zhang
44b4a0c9c5 [LLM] Correct prompt format of Yi, Llama2 and Qwen in generate.py (#9786) 
* correct prompt format of Yi

* correct prompt format of llama2 in cpu generate.py

* correct prompt format of Qwen in GPU example
 2023-12-26 16:57:55 +08:00
Xiangyu Tian
0ea842231e [LLM] vLLM: Add api_server entrypoint (#9783) 
Add vllm.entrypoints.api_server for benchmark_serving.py in vllm.
 2023-12-26 16:03:57 +08:00
dingbaorong
64d05e581c add peak gpu mem stats in transformer_int4_gpu (#9766) 
* add peak gpu mem stats in transformer_int4_gpu

* address weiguang's comments
 2023-12-26 15:38:28 +08:00
Ziteng Zhang
87b4100054 [LLM] Support Yi model in chat.py (#9778) 
* Suppot Yi model

* code style& add reference link
 2023-12-26 10:03:39 +08:00
Ruonan Wang
11d883301b LLM: fix wrong batch output caused by flash attention (#9780) 
* fix

* meet code review

* move batch size check to the beginning

* move qlen check inside function

* meet code review
 2023-12-26 09:41:27 +08:00
Heyang Sun
66e286a73d Support for Mixtral AWQ (#9775) 
* Support for Mixtral AWQ

* Update README.md

* Update README.md

* Update awq_config.py

* Update README.md

* Update README.md
 2023-12-25 16:08:09 +08:00
Ruonan Wang
1917bbe626 LLM: fix BF16Linear related training & inference issue (#9755) 
* fix bf16 related issue

* fix

* update based on comment & add arc lora script

* update readme

* update based on comment

* update based on comment

* update

* force to bf16

* fix style

* move check input dtype into function

* update convert

* meet code review

* meet code review

* update merged model to support new training_mode api

* fix typo
 2023-12-25 14:49:30 +08:00
Xiangyu Tian
30dab36f76 [LLM] vLLM: Fix kv cache init (#9771) 
Fix kv cache init
 2023-12-25 14:17:06 +08:00
Yina Chen
449b387125 Support relora in bigdl-llm (#9687) 
* init

* fix style

* update

* support resume & update readme

* update

* update

* remove important

* add training mode

* meet comments
 2023-12-25 14:04:28 +08:00
Shaojun Liu
b6222404b8 bigdl-llm stable version: let the perf test fail if the difference between perf and baseline is greater than 5% (#9750) 
* test

* test

* test

* update

* revert
 2023-12-25 13:47:11 +08:00
Ziteng Zhang
986f65cea9 [LLM] Add trust_remote_code for local renamed model in bigdl_llm_model.py (#9762) 2023-12-25 11:31:14 +08:00
Yishuo Wang
be13b162fe add codeshell example (#9743) 2023-12-25 10:54:01 +08:00
Guancheng Fu
daf536fb2d vLLM: Apply attention optimizations for selective batching (#9758) 
* fuse_rope for prefil

* apply kv_cache optimizations

* apply fast_decoding_path

* Re-enable kv_cache optimizations for prefill

* reduce KV_CACHE_ALLOC_BLOCK for selective_batching
 2023-12-25 10:29:31 +08:00
binbin Deng
ed8ed76d4f LLM: update deepspeed autotp usage (#9733) 2023-12-25 09:41:14 +08:00
Yuwen Hu
02436c6cce [LLM] Enable more long context in-out pairs for iGPU perf tests (#9765) 
* Add test for 1024-128 and enable more tests for 512-64

* Fix date in results csv name to the time when the performance is triggered

* Small fix

* Small fix

* further fixes
 2023-12-22 18:18:23 +08:00
Chen, Zhentao
7fd7c37e1b Enable fp8e5 harness (#9761) 
* fix precision format like fp8e5

* match fp8_e5m2
 2023-12-22 16:59:48 +08:00
Qiyuan Gong
4c487313f2 Revert "[LLM] IPEX auto importer turn on by default for XPU (#9730)" (#9759) 
This reverts commit 0284801fbd.
 2023-12-22 16:38:24 +08:00
Qiyuan Gong
0284801fbd [LLM] IPEX auto importer turn on by default for XPU (#9730) 
* Set BIGDL_IMPORT_IPEX default to true, i.e., auto import IPEX for XPU.
* Remove import intel_extension_for_pytorch as ipex from GPU example.
* Add support for bigdl-core-xe-21.
 2023-12-22 16:20:32 +08:00
Chen, Zhentao
86a69e289c fix harness runner label of manual trigger (#9754) 
* fix runner

* update golden
 2023-12-22 15:09:22 +08:00
Guancheng Fu
fdf93c9267 Implement selective batching for vLLM (#9659) 
* add control to load hf model

* finish initial version of selective_batching

* temp

* finish

* Remove print statement

* fix error

* Apply yang's optimization

* a version that works

* We need to check kv_cache passed in, this could be an error. TODO: add fast decoding path

* format

* temp solution: not batching prefill requests

* a version that works for prefill batching

* format

* a solid version: works normally

* a temp version

* Solid version: remove redundant functions

* fix format

* format

* solid: add option to enable selective_batching

* remove logic for using transformer models

* format

* format

* solid: enable argument VLLM_ENABLE_SELECTIVE_BATCHING

* format

* finish

* format
 2023-12-22 13:45:46 +08:00
Ruonan Wang
2f36769208 LLM: bigdl-llm lora support & lora example (#9740) 
* lora support and single card example

* support multi-card, refactor code

* fix model id and style

* remove torch patch, add two new class for bf16, update example

* fix style

* change to training_mode

* small fix

* add more info in help

* fixstyle, update readme

* fix ut

* fix ut

* Handling compatibility issues with default LoraConfig
 2023-12-22 11:05:39 +08:00
SONG Ge
ba0b939579 [LLM] Support transformers-v4.36.0 on mistral model (#9744) 
* add support transformers-v4.36.0 on mistral model

* python/llm/src/bigdl/llm/transformers/models/mistral.py

* make the redundant implementation as utils

* fix code style

* fix

* fix style

* update with utils enough_kv_room
 2023-12-22 09:59:27 +08:00
Xin Qiu
e36111e713 mixstral fused qkv and rope (#9724) 
* mixstral fused qkv and rope

* fix and clean

* fix style

* update

* update

* fix

* update

* fix
 2023-12-22 09:26:35 +08:00
Jiao Wang
e4f6e43675 safetenor to false (#9728) 2023-12-21 14:41:51 -08:00
Shaojun Liu
bb52239e0a bigdl-llm stable version release & test (#9732) 
* stable version test

* trigger spr test

* update

* trigger

* test

* test

* test

* test

* test

* refine

* release linux first
 2023-12-21 22:55:33 +08:00
WeiguangHan
d4d2ccdd9d LLM: remove startcorder-15.5b (#9748) 2023-12-21 18:52:52 +08:00
WeiguangHan
474c099559 LLM: using separate threads to do inference (#9727) 
* using separate threads to do inference

* resolve some comments

* resolve some comments

* revert llm_performance_tests.yml file
 2023-12-21 17:56:43 +08:00
Yishuo Wang
426660b88e simplify qwen attention (#9747) 2023-12-21 17:53:29 +08:00
Wang, Jian4
984697afe2 LLM: Add bloom gguf support (#9734) 
* init

* update bloom add merges

* update

* update readme

* update for llama error

* update
 2023-12-21 14:06:25 +08:00
Heyang Sun
df775cf316 fix python style (#9742) 
* fix python style

* fix

* fix
 2023-12-21 11:25:05 +08:00
Chen, Zhentao
b06a3146c8 Fix 70b oom (#9738) 
* add default value to bigdl llm

* fix model oom
 2023-12-21 10:40:52 +08:00
Xin Qiu
6c3e698bf1 mistral decoding_fast_path and fused mlp (#9714) 
* mistral decoding_fast_path and fused mlp

* meet code review
 2023-12-21 10:11:37 +08:00
Heyang Sun
d157f623b6 Load Mixtral gguf in a block-wise way (#9725) 
* Load Mixtral gguf in a block-wise way

* refine
 2023-12-21 10:03:23 +08:00
WeiguangHan
34bb804189 LLM: check csv and its corresponding yaml file (#9702) 
* LLM: check csv and its corresponding yaml file

* run PR arc perf test

* modify the name of some variables

* execute the check results script in right place

* use cp to replace mv command

* resolve some comments

* resolve more comments

* revert the llm_performance_test.yaml file
 2023-12-21 09:54:33 +08:00
Zhao Changmin
4bda975a3e LLM: Align lowbit model config (#9735) 
* align lowbit model config
 2023-12-21 09:48:58 +08:00
Wang, Jian4
e1e921f425 LLM: gguf other model using dtype (#9729) 2023-12-21 09:33:40 +08:00
Yishuo Wang
13ea6330bd optimize qwen rope (#9737) 2023-12-20 17:34:34 +08:00
Ziteng Zhang
4c032a433e [LLM] Add glibc checker (#9624) 
* Add glibc checker
* Add env BIGDL_GLIBC_CHECK to control glibc checker. The default is false, i.e., don't check.
 2023-12-20 16:52:43 +08:00
Yina Chen
cd652a1710 Support fp8 e5m2 on arc (#9711) 
* init

* fix style

* update

* fix style

* update
 2023-12-20 16:26:17 +08:00
Yishuo Wang
e54c428d30 add bf16/fp16 fuse mlp support (#9726) 2023-12-20 10:40:45 +08:00
Heyang Sun
612651cb5d fix typo (#9723) 2023-12-20 09:41:59 +08:00
WeiguangHan
3aa8b66bc3 LLM: remove starcoder-15.5b model temporarily (#9720) 2023-12-19 20:14:46 +08:00
Yishuo Wang
522cf5ed82 [LLM] Improve chatglm2/3 rest token performance with long context (#9716) 2023-12-19 17:29:38 +08:00
Yishuo Wang
f2e6abb563 fix mlp batch size check (#9718) 2023-12-19 14:22:22 +08:00
Heyang Sun
1fa7793fc0 Load Mixtral GGUF Model (#9690) 
* Load Mixtral GGUF Model

* refactor

* fix empty tensor when to cpu

* update gpu and cpu readmes

* add dtype when set tensor into module
 2023-12-19 13:54:38 +08:00
Qiyuan Gong
d0a3095b97 [LLM] IPEX auto importer (#9706) 
* IPEX auto importer and get_ipex_version.
* Add BIGDL_IMPORT_IPEX to control auto import, default is false.
 2023-12-19 13:39:38 +08:00
Yang Wang
f4fb58d99c fusing qkv project and rope (#9612) 
* Try fusing qkv project and rope

* add fused mlp

* fuse append cache

* fix style and clean up code

* clean up
 2023-12-18 16:45:00 -08:00
Kai Huang
4c112ee70c Rename qwen in model name for arc perf test (#9712) 2023-12-18 20:34:31 +08:00
Cengguang Zhang
4d22add4af LLM: fix qwen efficiency issue in perf-test. 2023-12-18 18:32:54 +08:00
Ruonan Wang
8ed89557e5 LLM: add mlp optimization of mixtral (#9709) 2023-12-18 16:59:52 +08:00
Chen, Zhentao
b3647507c0 Fix harness workflow (#9704) 
* error when larger than 0.001

* fix env setup

* fix typo

* fix typo
 2023-12-18 15:42:10 +08:00
binbin Deng
12df70953e LLM: add resume_from_checkpoint related section (#9705) 2023-12-18 12:27:02 +08:00
Xin Qiu
320110d158 handle empty fused norm result (#9688) 
* handle empty fused norm result

* remove fast_rms_norm

* fix style
 2023-12-18 09:56:11 +08:00
Ziteng Zhang
67cc155771 [LLM] Correct chat format of llama and add llama_stream_chat in chat.py 
* correct chat format of llama
* add llama_stream_chat
 2023-12-15 16:36:46 +08:00
Ziteng Zhang
0d41b7ba7b [LLM] Correct chat format & add stop words for chatglm3 in chat.py 
* correct chat format of chatglm3
* correct stop words of chatglm3
 2023-12-15 16:35:17 +08:00
Ziteng Zhang
d57efd8eb9 [LM] Add stop_word for Qwen model and correct qwen chat format in chat.py (#9642) 
* add stop words list for qwen

* change qwen chat format
 2023-12-15 14:53:58 +08:00
SONG Ge
d5b81af7bd Support mixtral attention optimization on transformers-v4.36.0 (#9674) 
* add example code to support mistral/mixtral attention on transformers v4.36.0

* update

* style fix

* add update for seen-tokens

* support mixtral

* rm mistral change

* small fix

* add more comments and remove use_cache part

---------

Co-authored-by: plusbang <binbin1.deng@intel.com>
 2023-12-15 14:30:23 +08:00
Cengguang Zhang
adbef56001 LLM: update qwen attention forward. (#9695) 
* feat: update qwen attention forward.

* fix: style.
 2023-12-15 14:06:15 +08:00
Wang, Jian4
b8437a1c1e LLM: Add gguf mistral model support (#9691) 
* add mistral support

* need to upgrade transformers version

* update
 2023-12-15 13:37:39 +08:00
Wang, Jian4
496bb2e845 LLM: Support load BaiChuan model family gguf model (#9685) 
* support baichuan model family gguf model

* update gguf generate.py

* add verify models

* add support model_family

* update

* update style

* update type

* update readme

* update

* remove support model_family
 2023-12-15 13:34:33 +08:00
Lilac09
3afed99216 fix path issue (#9696) 2023-12-15 11:21:49 +08:00
Jason Dai
37f509bb95 Update readme (#9692) 2023-12-14 19:50:21 +08:00
WeiguangHan
1f0245039d LLM: check the final csv results for arc perf test (#9684) 
* LLM: check the final csv results for arc perf test

* delete useless python script

* change threshold

* revert the llm_performance_tests.yml
 2023-12-14 19:46:08 +08:00
Yishuo Wang
9a330bfc2b fix fuse mlp when using q5_0 or fp8 (#9689) 2023-12-14 16:16:05 +08:00
Yuwen Hu
82ac2dbf55 [LLM] Small fixes for win igpu test for ipex 2.1 (#9686) 
* Fixes to install for igpu performance tests

* Small update for core performance tests model lists
 2023-12-14 15:39:51 +08:00
WeiguangHan
3e8d198b57 LLM: add eval func (#9662) 
* Add eval func

* add left eval
 2023-12-14 14:59:02 +08:00
Ziteng Zhang
21c7503a42 [LLM] Correct prompt format of Qwen in generate.py (#9678) 
* Change qwen prompt format to chatml
 2023-12-14 14:01:30 +08:00
Qiyuan Gong
223c9622f7 [LLM] Mixtral CPU examples (#9673) 
* Mixtral CPU PyTorch and hugging face examples, based on #9661 and #9671
 2023-12-14 10:35:11 +08:00
Xin Qiu
5e46e0e5af fix baichuan2-7b 1st token performance regression on xpu (#9683) 
* fix baichuan2-7b 1st token performance regression

* add comments

* fix style
 2023-12-14 09:58:32 +08:00
ZehuaCao
877229f3be [LLM]Add Yi-34B-AWQ to verified AWQ model. (#9676) 
* verfiy Yi-34B-AWQ

* update
 2023-12-14 09:55:47 +08:00
binbin Deng
68a4be762f remove disco mixtral, update oneapi version (#9671) 2023-12-13 23:24:59 +08:00
Ruonan Wang
1456d30765 LLM: add dot to option name in setup (#9682) 2023-12-13 20:57:27 +08:00
Yuwen Hu
cbdd49f229 [LLM] win igpu performance for ipex 2.1 and oneapi 2024.0 (#9679) 
* Change igpu win tests for ipex 2.1 and oneapi 2024.0

* Qwen model repo id updates; updates model list for 512-64

* Add .eval for win igpu all-in-one benchmark for best performance
 2023-12-13 18:52:29 +08:00
Mingyu Wei
16febc949c [LLM] Add exclude option in all-in-one performance test (#9632) 
* add exclude option in all-in-one perf test

* update arc-perf-test.yaml

* Exclude in_out_pairs in main function

* fix some bugs

* address Kai's comments

* define excludes at the beginning

* add bloomz:2048 to exclude
 2023-12-13 18:13:06 +08:00
Ruonan Wang
9b9cd51de1 LLM: update setup to provide new install option to support ipex 2.1 & oneapi 2024 (#9647) 
* update setup

* default to 2.0 now

* meet code review
 2023-12-13 17:31:56 +08:00
Yishuo Wang
09ca540f9b use fuse mlp in qwen (#9672) 2023-12-13 17:20:08 +08:00
Ruonan Wang
c7741c4e84 LLM: update moe block convert to optimize rest token latency of Mixtral (#9669) 
* update moe block convert

* further accelerate final_hidden_states

* fix style

* fix style
 2023-12-13 16:17:06 +08:00
ZehuaCao
503880809c verfiy codeLlama (#9668) 2023-12-13 15:39:31 +08:00
Xiangyu Tian
1c6499e880 [LLM] vLLM: Support Mixtral Model (#9670) 
Add Mixtral support for BigDL vLLM.
 2023-12-13 14:44:47 +08:00
Ruonan Wang
dc5b1d7e9d LLM: integrate sdp kernel for FP16 rest token inference on GPU [DG2/ATSM] (#9633) 
* integrate sdp

* update api

* fix style

* meet code review

* fix

* distinguish mtl from arc

* small fix
 2023-12-13 11:29:57 +08:00
Qiyuan Gong
5b0e7e308c [LLM] Add support for empty activation (#9664) 
* Add support for empty activation, e.g., [0, 4096]. Empty activation is allowed by PyTorch.
* Add comments.
 2023-12-13 11:07:45 +08:00
SONG Ge
284e7697b1 [LLM] Optimize ChatGLM2 kv_cache to support beam_search on ARC (#9579) 
* optimize kv_cache to support beam_search on Arc

* correctness test update

* fix query_length issue

* simplify implementation

* only enable the optimization on gpu device

* limit the beam_search support only enabled with gpu device and batch_size > 1

* add comments for beam_search case and revert ut change

* meet comments

* add more comments to describe the differece between multi-cases
 2023-12-13 11:02:14 +08:00
Heyang Sun
c64e2248ef fix str returned by get_int_from_str rather than expected int (#9667) 2023-12-13 11:01:21 +08:00
binbin Deng
bf1bcf4a14 add official Mixtral model support (#9663) 2023-12-12 22:27:07 +08:00
Ziteng Zhang
8931f2eb62 [LLM] Fix transformer qwen size mismatch and rename causal_mask (#9655) 
* Fix size mismatching caused by context_layer
* Change registered_causal_mask to causal_mask
 2023-12-12 20:57:40 +08:00
binbin Deng
2fe38b4b9b LLM: add mixtral GPU examples (#9661) 2023-12-12 20:26:36 +08:00
Yuwen Hu
968d99e6f5 Remove empty cache between each iteration of generation (#9660) 2023-12-12 17:24:06 +08:00
Xin Qiu
0e639b920f disable test_optimized_model.py temporarily due to out of memory on A730M(pr validation machine) (#9658) 
* disable test_optimized_model.py

* disable seq2seq
 2023-12-12 17:13:52 +08:00
binbin Deng
59ce86d292 LLM: support optimize_model=True for Mixtral model (#9657) 2023-12-12 16:41:26 +08:00
Yuwen Hu
d272b6dc47 [LLM] Enable generation of html again for win igpu tests (#9652) 
* Enable generation of html again and comment out rwkv for 32-512 as it is not very stable

* Small fix
 2023-12-11 19:15:17 +08:00
WeiguangHan
afa895877c LLM: fix the issue that may generate blank html (#9650) 
* LLM: fix the issue that may generate blank html

* reslove some comments
 2023-12-11 19:14:57 +08:00
ZehuaCao
45721f3473 verfiy llava (#9649) 2023-12-11 14:26:05 +08:00
Heyang Sun
9f02f96160 [LLM] support for Yi AWQ model (#9648) 2023-12-11 14:07:34 +08:00
Xin Qiu
82255f9726 Enable fused layernorm (#9614) 
* bloom layernorm

* fix

* layernorm

* fix

* fix

* fix

* style fix

* fix

* replace nn.LayerNorm
 2023-12-11 09:26:13 +08:00
Yuwen Hu
894d0aaf5e [LLM] iGPU win perf test reorg based on in-out pairs (#9645) 
* trigger pr temparorily

* Saparate benchmark run for win igpu based in in-out pairs

* Rename fix

* Test workflow

* Small fix

* Skip generation of html for now

* Change back to nightly triggered
 2023-12-08 20:46:40 +08:00
Chen, Zhentao
972cdb9992 gsm8k OOM workaround (#9597) 
* update bigdl_llm.py

* update the installation of harness

* fix partial function

* import ipex

* force seq len in decrease order

* put func outside class

* move comments

* default 'trust_remote_code' as True

* Update llm-harness-evaluation.yml
 2023-12-08 18:47:25 +08:00
WeiguangHan
1ff4bc43a6 degrade pandas version (#9643) 2023-12-08 17:44:51 +08:00
Yina Chen
70f5e7bf0d Support peft LoraConfig (#9636) 
* support peft loraconfig

* use testcase to test

* fix style

* meet comments
 2023-12-08 16:13:03 +08:00
Xin Qiu
0b6f29a7fc add fused rms norm for Yi and Qwen (#9640) 2023-12-08 16:04:38 +08:00
Xin Qiu
5636b0ba80 set new linear status (#9639) 2023-12-08 11:02:49 +08:00
binbin Deng
499100daf1 LLM: Add solution to fix oneccl related error (#9630) 2023-12-08 10:51:55 +08:00
ZehuaCao
6eca8a8bb5 update transformer version (#9631) 2023-12-08 09:36:00 +08:00
WeiguangHan
e9299adb3b LLM: Highlight some values in the html (#9635) 
* highlight some values in the html

* revert the llm_performance_tests.yml
 2023-12-07 19:02:41 +08:00
Yuwen Hu
6f34978b94 [LLM] Add more performance tests for win iGPU (more in-out pairs, RWKV model) (#9626) 
* Add supports for loading rwkv models using from_pretrained api

* Temporarily enable pr tests

* Add RWKV in tests and more in-out pairs

* Add rwkv for 512 tests

* Make iterations smaller

* Change back to nightly trigger
 2023-12-07 18:55:16 +08:00
Ruonan Wang
d9b0c01de3 LLM: fix unlora module in qlora finetune (#9621) 
* fix unlora module

* split train and inference
 2023-12-07 16:32:02 +08:00
Heyang Sun
3811cf43c9 [LLM] update AWQ documents (#9623) 
* [LLM] update AWQ and verified models' documents

* refine

* refine links

* refine
 2023-12-07 16:02:20 +08:00
Yishuo Wang
7319f2c227 use fused mlp in baichuan2 (#9620) 2023-12-07 15:50:57 +08:00
Xiangyu Tian
deee65785c [LLM] vLLM: Delete last_kv_cache before prefilling (#9619) 
Remove last_kv_cache before prefilling to reduce peak memory usage.
 2023-12-07 11:32:33 +08:00
Yuwen Hu
48b85593b3 Update all-in-one benchmark readme (#9618) 2023-12-07 10:32:09 +08:00
Xiangyu Tian
0327169b50 [LLM] vLLM: fix memory leak in prepare_kv_cache (#9616) 
Revert modification in prepare_kv_cache to fix memory leak.
 2023-12-07 10:08:18 +08:00
Xin Qiu
13d47955a8 use fused rms norm in chatglm2 and baichuan (#9613) 
* use fused rms norm in chatglm2 and baichuan

* style fix
 2023-12-07 09:21:41 +08:00
Jason Dai
51b668f229 Update GGUF readme (#9611) 2023-12-06 18:21:54 +08:00
dingbaorong
a7bc89b3a1 remove q4_1 in gguf example (#9610) 
* remove q4_1

* fixes
 2023-12-06 16:00:05 +08:00
Yina Chen
404e101ded QALora example (#9551) 
* Support qa-lora

* init

* update

* update

* update

* update

* update

* update merge

* update

* fix style & update scripts

* update

* address comments

* fix typo

* fix typo

---------

Co-authored-by: Yang Wang <yang3.wang@intel.com>
 2023-12-06 15:36:21 +08:00
Guancheng Fu
6978b2c316 [VLLM] Change padding patterns for vLLM & clean code (#9609) 
* optimize

* fix minor error

* optimizations

* fix style
 2023-12-06 15:27:26 +08:00
dingbaorong
89069d6173 Add gpu gguf example (#9603) 
* add gpu gguf example

* some fixes

* address kai's comments

* address json's comments
 2023-12-06 15:17:54 +08:00
Yuwen Hu
0e8f4020e5 Add traceback error output for win igpu test api in benchmark (#9607) 2023-12-06 14:35:16 +08:00
Ziteng Zhang
aeb77b2ab1 Add minimum Qwen model version (#9606) 2023-12-06 11:49:14 +08:00
Yuwen Hu
c998f5f2ba [LLM] iGPU long context tests (#9598) 
* Temp enable PR

* Enable tests for 256-64

* Try again 128-64

* Empty cache after each iteration for igpu benchmark scripts

* Try tests for 512

* change order for 512

* Skip chatglm3 and llama2 for now

* Separate tests for 512-64

* Small fix

* Further fixes

* Change back to nightly again
 2023-12-06 10:19:20 +08:00
Heyang Sun
4e70e33934 [LLM] code and document for distributed qlora (#9585) 
* [LLM] code and document for distributed qlora

* doc

* refine for gradient checkpoint

* refine

* Update alpaca_qlora_finetuning_cpu.py

* Update alpaca_qlora_finetuning_cpu.py

* Update alpaca_qlora_finetuning_cpu.py

* add link in doc
 2023-12-06 09:23:17 +08:00
Zheng, Yi
d154b38bf9 Add llama2 gpu low memory example (#9514) 
* Add low memory example

* Minor fixes

* Update readme.md
 2023-12-05 17:29:48 +08:00
Jason Dai
06febb5fa7 Update readme for FP8/FP4 inference examples (#9601) 2023-12-05 15:59:03 +08:00
dingbaorong
a66fbedd7e add gpu more data types example (#9592) 
* add gpu more data types example

* add int8
 2023-12-05 15:45:38 +08:00
Ziteng Zhang
65934c9f4f [LLM] Fix Qwen causal_mask and attention_mask size mismatching (#9600) 
* Fix #9582 , caused by Qwen modified modeling_qwen.py 7f62181c94 (d2h-049182)
 2023-12-05 15:15:54 +08:00
Jinyi Wan
b721138132 Add cpu and gpu examples for BlueLM (#9589) 
* Add cpu int4 example for BlueLM

* addexample optimize_model cpu for bluelm

* add example gpu int4 blueLM

* add example optimiza_model GPU for bluelm

* Fixing naming issues and BigDL package version.

* Fixing naming issues...

* Add BlueLM in README.md "Verified Models"
 2023-12-05 13:59:02 +08:00
Guancheng Fu
8b00653039 fix doc (#9599) 2023-12-05 13:49:31 +08:00
Qiyuan Gong
f211f136b6 Configurable TORCH_LINEAR_THRESHOLD from env (#9588) 
* Add TORCH_LINEAR_THRESHOLD from env (BIGDL_LLM_LINEAR_THRESHOLD)
* Change default to 512
 2023-12-05 13:19:47 +08:00
Yuwen Hu
1012507a40 [LLM] Fix performance tests (#9596) 
* Fix missing key for cpu_embedding

* Remove 512 as it stuck for now

* Small fix
 2023-12-05 10:59:28 +08:00
Chen, Zhentao
8c8a27ded7 Add harness summary job (#9457) 
* format yml

* add make_table_results

* add summary job

* add a job to print single result

* upload full directory
 2023-12-05 10:04:10 +08:00
Yuwen Hu
3f4ad97929 [LLM] Add performance tests for windows iGPU (#9584) 
* Add support for win gpu benchmark with peak gpu memory monitoring

* Add win igpu tests

* Small fix

* Forward outputs

* Small fix

* Test and small fixes

* Small fix

* Small fix and test

* Small fixes

* Add tests for 512-64 and change back to nightly tests

* Small fix
 2023-12-04 20:50:02 +08:00
Chen, Zhentao
9557aa9c21 Fix harness nightly (#9586) 
* update golden

* loose the restriction of diff

* only compare results when scheduled
 2023-12-04 11:45:00 +08:00
Xiangyu Tian
5c03651309 [LLM] vLLM: Add Preempt for scheduler (#9568) 
Implement Preempt_by_recompute method for vllm.
 2023-12-03 20:16:25 +08:00
Chen, Zhentao
cb228c70ea Add harness nightly (#9552) 
* modify output_path as a directory

* schedule nightly at 21 on Friday

* add tasks and models for nightly

* add accuracy regression

* comment out if to test

* mixed fp4

* for test

* add  missing delimiter

* remove comma

* fixed golden results

* add mixed 4 golden result

* add more options

* add mistral results

* get golden result of stable lm

* move nightly scripts and results to test folder

* add license

* add fp8 stable lm golden

* run on all available devices

* trigger only when ready for review

* fix new line

* update golden

* add mistral
 2023-12-01 14:16:35 +08:00
Chen, Zhentao
4d7d5d4c59 Add 3 leaderboard tasks (#9566) 
* update leaderboard map

* download model and dataset without overwritten

* fix task drop

* run on all available devices
 2023-12-01 14:01:14 +08:00
Wang, Jian4
ed0dc57c6e LLM: Add cpu qlora support other models guide (#9567) 
* use bf16 flag

* add using baichuan model

* update merge

* remove

* update
 2023-12-01 11:18:04 +08:00
Jason Dai
bda404fc8f Update readme (#9575) 2023-11-30 22:45:52 +08:00
Xin Qiu
69c49d21f5 use fused rms norm (#9572) 
* use fused rms norm

* meet code review
 2023-11-30 21:47:41 +08:00
Yishuo Wang
66f5b45f57 [LLM] add a llama2 gguf example (#9553) 2023-11-30 16:37:17 +08:00
Yishuo Wang
7f6465518a support loading llama tokenizer from gguf model (#9565) 2023-11-30 14:56:12 +08:00
Wang, Jian4
a0a80d232e LLM: Add qlora cpu distributed readme (#9561) 
* init readme

* add distributed guide

* update
 2023-11-30 13:42:30 +08:00
Chen, Zhentao
c8e0c2ed48 Fixed dumped logs in harness (#9549) 
* install transformers==4.34.0

* modify output_path as a directory

* add device and task to output dir parents
 2023-11-30 12:47:56 +08:00
Qiyuan Gong
d85a430a8c Uing bigdl-llm-init instead of bigdl-nano-init (#9558) 
* Replace `bigdl-nano-init` with `bigdl-llm-init`.
* Install `bigdl-llm` instead of `bigdl-nano`.
* Remove nano in README.
 2023-11-30 10:10:29 +08:00
Yuwen Hu
34503efa6a Fix cpu pinned embedding (#9556) 2023-11-29 18:27:56 +08:00
binbin Deng
4ff2ca9d0d LLM: fix loss error on Arc (#9550) 2023-11-29 15:16:18 +08:00
Yishuo Wang
65121c7997 support loading q4_1/q5_0/q5_1/q8_0 gguf model (#9546) 2023-11-29 14:40:37 +08:00
Wang, Jian4
b824754256 LLM: Update for cpu qlora mpirun (#9548) 2023-11-29 10:56:17 +08:00
Yuwen Hu
5f5ca38b74 [LLM Doc] Fix api doc rendering error (#9542) 
* Fix api rendering error

* Fix python style
 2023-11-29 09:17:09 +08:00
Yishuo Wang
a86c6e0b56 [LLM] support loading gguf model (#9544) 2023-11-28 15:51:15 +08:00
Xiangyu Tian
916c338772 fix bugs in vllm length check (#9543) 2023-11-28 11:09:54 +08:00
WeiguangHan
5098bc3544 LLM: enable previous models (#9505) 
* enable previous models

* test mistral model

* for test

* run models separately

* test all models

* for test

* revert the llm_performance_test.yaml
 2023-11-28 10:21:07 +08:00
Zhao Changmin
e7e0cd3b5e CPU Pinned embedding Layer (#9538) 
* CPU Pinned embedding
 2023-11-28 09:46:31 +08:00
Guancheng Fu
963a5c8d79 Add vLLM-XPU version's README/examples (#9536) 
* test

* test

* fix last kv cache

* add xpu readme

* remove numactl for xpu example

* fix link error

* update max_num_batched_tokens logic

* add explaination

* add xpu environement version requirement

* refine gpu memory

* fix

* fix style
 2023-11-28 09:44:03 +08:00
Guancheng Fu
b6c3520748 Remove xformers from vLLM-CPU (#9535) 2023-11-27 11:21:25 +08:00
binbin Deng
2b9c7d2a59 LLM: quick fix alpaca qlora finetuning script (#9534) 2023-11-27 11:04:27 +08:00
Yuwen Hu
11fa3de290 Add sutup support of win gpu for bigdl-llm (#9512) 2023-11-24 17:49:21 +08:00
Chen, Zhentao
45820cf3b9 add optimize model option (#9530) 2023-11-24 17:10:49 +08:00
binbin Deng
6bec0faea5 LLM: support Mistral AWQ models (#9520) 2023-11-24 16:20:22 +08:00
Ruonan Wang
914a5a5a27 LLM: fix abnormal Mistral GPU accuracy by updating rms_norm (#9529) 2023-11-24 15:37:50 +08:00
SONG Ge
3d24823cda hot-fix mistral kv_cache (#9528) 2023-11-24 14:33:04 +08:00
Zhao Changmin
42b7a16bc5 Replace torch.bmm with safe_bmm (#9519) 
* replace bmm with safe one

* rename args and deprecated warning
 2023-11-24 12:16:48 +08:00
Jason Dai
b3178d449f Update README.md (#9525) 2023-11-23 21:45:20 +08:00
Jason Dai
82898a4203 Update GPU example README (#9524) 2023-11-23 21:20:26 +08:00
Jason Dai
064848028f Update README.md (#9523) 2023-11-23 21:16:21 +08:00
Ruonan Wang
b63aae8a8e LLM: add flash attention support for llama (#9518) 
* add initial flash attention for llama

* accelerate fp32 first token by changing to fp16 in advance

* support fp32
 2023-11-23 18:40:18 +08:00
Guancheng Fu
bf579507c2 Integrate vllm (#9310) 
* done

* Rename structure

* add models

* Add structure/sampling_params,sequence

* add input_metadata

* add outputs

* Add policy,logger

* add and update

* add parallelconfig back

* core/scheduler.py

* Add llm_engine.py

* Add async_llm_engine.py

* Add tested entrypoint

* fix minor error

* Fix everything

* fix kv cache view

* fix

* fix

* fix

* format&refine

* remove logger from repo

* try to add token latency

* remove logger

* Refine config.py

* finish worker.py

* delete utils.py

* add license

* refine

* refine sequence.py

* remove sampling_params.py

* finish

* add license

* format

* add license

* refine

* refine

* Refine line too long

* remove exception

* so dumb style-check

* refine

* refine

* refine

* refine

* refine

* refine

* add README

* refine README

* add warning instead error

* fix padding

* add license

* format

* format

* format fix

* Refine vllm dependency (#1)

vllm dependency clear

* fix licence

* fix format

* fix format

* fix

* adapt LLM engine

* fix

* add license

* fix format

* fix

* Moving README.md to the correct position

* Fix readme.md

* done

* guide for adding models

* fix

* Fix README.md

* Add new model readme

* remove ray-logic

* refactor arg_utils.py

* remove distributed_init_method logic

* refactor entrypoints

* refactor input_metadata

* refactor model_loader

* refactor utils.py

* refactor models

* fix api server

* remove vllm.stucture

* revert by txy 1120

* remove utils

* format

* fix license

* add bigdl model

* Refer to a specfic commit

* Change code base

* add comments

* add async_llm_engine comment

* refine

* formatted

* add worker comments

* add comments

* add comments

* fix style

* add changes

---------

Co-authored-by: xiangyuT <xiangyu.tian@intel.com>
Co-authored-by: Xiangyu Tian <109123695+xiangyuT@users.noreply.github.com>
Co-authored-by: leonardozcm <leonardo1997zcm@gmail.com>
 2023-11-23 16:46:45 +08:00
Heyang Sun
48fbb1eb94 support ccl (MPI) distributed mode in alpaca_qlora_finetuning_cpu (#9507) 2023-11-23 10:58:09 +08:00
Qiyuan Gong
0f0c6bb631 [LLM] Fix Qwen registered_causal_mask is None (#9513) 
* Add registered_causal_mask init based on 2abd8e5777.
 2023-11-23 09:28:04 +08:00
Heyang Sun
11fa5a8a0e Fix QLoRA CPU dispatch_model issue about accelerate (#9506) 2023-11-23 08:41:25 +08:00
Heyang Sun
1453046938 install bigdl-llm in deepspeed cpu inference example (#9508) 2023-11-23 08:39:21 +08:00
binbin Deng
86743fb57b LLM: fix transformers version in CPU finetuning example (#9511) 2023-11-22 15:53:07 +08:00
binbin Deng
1a2129221d LLM: support resume from checkpoint in Alpaca QLoRA (#9502) 2023-11-22 13:49:14 +08:00
Ruonan Wang
139e98aa18 LLM: quick fix benchmark (#9509) 2023-11-22 10:19:57 +08:00
WeiguangHan
c2aeb4d1e8 del model after test (#9504) 2023-11-21 18:41:50 +08:00
Ruonan Wang
076d106ef5 LLM: GPU QLoRA update to bf16 to accelerate gradient checkpointing (#9499) 
* update to bf16 to accelerate gradient checkpoint

* add utils and fix ut
 2023-11-21 17:08:36 +08:00
Cheen Hau, 俊豪
3e39828420 Update all in one benchmark readme (#9496) 
* Add gperftools install to all in one benchmark readme

* Update readme
 2023-11-21 14:57:16 +08:00
binbin Deng
b7ae572ac3 LLM: update Alpaca QLoRA finetuning example on GPU (#9492) 2023-11-21 14:22:19 +08:00
Wang, Jian4
c5cb3ab82e LLM : Add CPU alpaca qlora example (#9469) 
* init

* update xpu to cpu

* update

* update readme

* update example

* update

* add refer

* add guide to train different datasets

* update readme

* update
 2023-11-21 09:19:58 +08:00
binbin Deng
96fd26759c LLM: fix QLoRA finetuning example on CPU (#9489) 2023-11-20 14:31:24 +08:00
Xin Qiu
50b01058f1 enable new q4_1 (#9479) 2023-11-17 14:58:57 +08:00
binbin Deng
3dac21ac7b LLM: add more example usages about alpaca qlora on different hardware (#9458) 2023-11-17 09:56:43 +08:00
Heyang Sun
921b263d6a update deepspeed install and run guide in README (#9441) 2023-11-17 09:11:39 +08:00
Zhao Changmin
30abd304a7 LLM: Fix baichuan pre-normalize model tensor assigning issue when loading (#9481) 
* No need to normalized when loading
 2023-11-16 21:57:28 +08:00
WeiguangHan
bc06bec90e LLM: modify the script to generate html results more accurately (#9445) 
* modify the script to generate html results more accurately

* resolve some comments

* revert some codes
 2023-11-16 19:50:23 +08:00
Ruonan Wang
c0ef70df02 llm: quick fix of fast_rms_norm (#9480) 2023-11-16 14:42:16 +08:00
Yina Chen
d5263e6681 Add awq load support (#9453) 
* Support directly loading GPTQ models from huggingface

* fix style

* fix tests

* change example structure

* address comments

* fix style

* init

* address comments

* add examples

* fix style

* fix style

* fix style

* fix style

* update

* remove

* meet comments

* fix style

---------

Co-authored-by: Yang Wang <yang3.wang@intel.com>
 2023-11-16 14:06:25 +08:00