binbin Deng
5ee6f97d6f
[NPU L0] Add layernorm weight as const / input setting ( #12322 )
2024-11-04 15:46:24 +08:00
Kai Huang
c8679ad592
Qwen layernorm as input ( #12309 )
...
* qwen layernorm as input
* add group size
2024-11-04 09:51:15 +08:00
Ch1y0q
48123af463
add npu_group_size for transformers_int4_npu_win in all-in-one benchmark api ( #12316 )
...
* add `npu_group_size` for `transformers_int4_npu_win`
small bugfix
* update
2024-11-01 18:44:27 +08:00
binbin Deng
f53bb4ea0b
[NPU L0] Update 1st token generation ( #12314 )
2024-11-01 17:02:07 +08:00
binbin Deng
d409d9d0eb
[NPU L0] Update streaming mode of example ( #12312 )
2024-11-01 15:38:10 +08:00
Yina Chen
05c5d0267a
[NPU] Llama2 prefill use ov sdp ( #12310 )
...
* prefill use sdp
* add param
* update
* fix style
* fix style
* meet comments
2024-11-01 11:05:20 +08:00
binbin Deng
eda764909c
Add minicpm-2b in L0 pipeline ( #12308 )
2024-11-01 09:30:01 +08:00
Yishuo Wang
b9853f98b3
fix qwen2 attention_mask slice ( #12307 )
2024-10-31 17:00:05 +08:00
binbin Deng
4892df61c9
Add qwen2-1.5b in l0 pipeline example ( #12306 )
2024-10-31 16:44:25 +08:00
Xin Qiu
97a0f7fd35
Codegeex support ( #12303 )
...
* new codegeex attn
* use kv cache
* add compress/quantize kv
* remove compress/quantize kv
* fix style check
* fix style
* fix codegeex
2024-10-31 15:28:56 +08:00
Yishuo Wang
72605c7016
fix llama3.1/3.2 quantize kv check ( #12302 )
2024-10-31 11:55:07 +08:00
Kai Huang
416c19165c
Add Qwen pipeline and example ( #12292 )
...
* support qwen pipeline
* update error msg
* style
* meet review
* minor
2024-10-31 11:25:25 +08:00
Yina Chen
0763268e4c
[NPU]Qwen2 groupwise performance opt ( #12299 )
...
* qwen2 gw performance opt
* remove debug
2024-10-30 17:40:21 +08:00
binbin Deng
41b8064554
Support minicpm-1B in level0 pipeline ( #12297 )
2024-10-30 17:21:47 +08:00
Jinhe
46d8300f6b
bugfix for qlora finetuning on GPU ( #12298 )
...
* bugfix for qlora 100 step error
* indent fix
* annotation fix
2024-10-30 16:54:10 +08:00
Yina Chen
70037ad55f
Groupwise prefill optimization ( #12291 )
...
* except lm_head
* remove
* support gw lm_head
* update
* fix
* remove run.bat
* fix style
* support llama3
* slice -> split
* remove debug
* fix style
* add dpu
2024-10-30 14:59:45 +08:00
Yishuo Wang
540eaeb12c
refactor attention_softmax ( #12295 )
2024-10-30 13:20:50 +08:00
Ruonan Wang
2b2cb9c693
[NPU pipeline] Support save & load and update examples ( #12293 )
...
* support save & load, update llama examples
* update baichuan2 example
* update readme
2024-10-30 10:02:00 +08:00
Yuwen Hu
5a15098835
Initial support for quantized forward on CPU when quantization_group_size=0 ( #12282 )
...
* Initial support for quantized forward on CPU when quantization_group_size=0
* Style fix
* Style fix
* Small fix
* Small fix
2024-10-29 19:40:17 +08:00
binbin Deng
3feb58d1e4
Support baichuan2 for level0 pipeline ( #12289 )
2024-10-29 19:24:16 +08:00
Zhao Changmin
546f455e8e
Patch sdpa check function in specific module attributes table ( #12285 )
2024-10-29 18:41:09 +08:00
Ruonan Wang
821b0033ed
[NPU L0] update layernorm & code refactor ( #12287 )
...
* update layernorm & code refactor
* fix style
* add common utils
* change to Pool()
* remove print
2024-10-29 15:01:45 +08:00
Yina Chen
4467645088
[NPU] Support l0 Llama groupwise ( #12276 )
...
* except lm_head
* remove
* support gw lm_head
* update
* fix
* remove run.bat
* fix style
* support llama3
2024-10-28 17:06:55 +08:00
Ruonan Wang
3fe2ea3081
[NPU] Reuse prefill of acc lib for pipeline ( #12279 )
...
* first commit
* update example
* fix style
* update example
* embedding as const
* fix generate
* code refactor
* meet code review
* fix style
* change max_output_len to max_context_len
* fix all-in-one
* fix example
* add check for new tokens
2024-10-28 16:05:49 +08:00
SONG Ge
08cb065370
hot-fix redundant import funasr ( #12277 )
2024-10-25 19:40:39 +08:00
SONG Ge
a0c6432899
[NPU] Add support for loading a FunASR model ( #12073 )
...
* add support for loading funasr model
* add initial support for paraformer-encoder
* add npu ops impl
* add encoder-decoder npu pipeline
* move paraformer encoders prefix 30 layers to npu and keep the rest layers on cpu
2024-10-25 17:22:01 +08:00
Yuwen Hu
43b25a2fe7
Fix llama 3.2 vision on LNL ( #12264 )
...
* Fix llama 3.2 vision on LNL
* Small fix
2024-10-25 16:23:31 +08:00
Ruonan Wang
ae57e23e4f
fix incompatibility between llama GW & llama pipeline ( #12267 )
...
* fix
* fix
2024-10-25 10:31:44 +08:00
Yina Chen
b5e663854b
[NPU] Support llama groupwise ( #12260 )
...
* support llama gw
* support llama gw lm_head
* fix style
* remove unused code
2024-10-24 18:06:45 +08:00
Xin Qiu
39c9d1de52
fix code geex ( #12261 )
2024-10-24 14:34:01 +08:00
Yishuo Wang
f3a2b20e6b
Optimize gpt2 ( #12259 )
2024-10-24 13:44:24 +08:00
Ruonan Wang
821fd96367
Initial integrate our L0 Llama impl into ipex-llm ( #12255 )
...
* temp save
* initial support
* fix
* simplify code
* fix style
* fix example
* make default value of pipeline as False
2024-10-24 09:49:27 +08:00
Yishuo Wang
cacc891962
Fix PR validation ( #12253 )
2024-10-23 18:10:47 +08:00
binbin Deng
b685cf4349
Fix npu group size setting of optimize_model=False ( #12256 )
2024-10-23 17:53:54 +08:00
binbin Deng
567b77a76b
Support IR and blob format for llama level0 pipeline ( #12251 )
2024-10-23 16:02:35 +08:00
Yishuo Wang
578aef245d
Fix models auto choose SdpaAttention with ipex 2.3 ( #12252 )
2024-10-23 15:33:45 +08:00
Yishuo Wang
88dc120a4c
fix fp16 linear ( #12250 )
2024-10-23 14:35:19 +08:00
Yina Chen
e8cf7f32f5
npu gw small fix ( #12249 )
2024-10-23 14:26:01 +08:00
Yina Chen
e37f951cce
[NPU] Groupwise ( #12241 )
...
* dq divide
* fix
* support attn divide
* update qwen2 7b
* divide down_proj & other linear
* use concat & reduce sum
* support scale after
* support qwen2
* w/ mm
* update reshape
* spda
* split
* split 2+
* update
* lm head-> 28
* no scale
* update
* update
* update
* fix style
* fix style
* to split linear
* update
* update code
* address comments
* fix style & remove redundant code & revert benchmark scripts
* fix style & remove code
* update save & load
---------
Co-authored-by: Yang Wang <yang3.wang@intel.com>
2024-10-23 14:10:58 +08:00
Yina Chen
ec465fbcd7
Add lookup generate in load_low_bit ( #12243 )
...
* add lookup generate in load_low_bit
* update comment
2024-10-22 15:51:52 +08:00
Yuwen Hu
b3df47486d
Fix Gemma 2 on LNL ( #12240 )
...
* Fix gemma 2 on LNL
* Python style fix
2024-10-21 18:25:53 +08:00
Yishuo Wang
9ea694484d
refactor ot remove old rope usage ( #12224 )
2024-10-17 17:06:09 +08:00
Yishuo Wang
324bcb057e
refactor to reduce old rope usage ( #12219 )
2024-10-17 14:45:09 +08:00
Yishuo Wang
a4a758656a
refactor gemma to reduce old fuse rope usage ( #12215 )
2024-10-16 17:40:28 +08:00
Yishuo Wang
9104a168f6
refactor phi-2 to reduce old fuse rope usage ( #12214 )
2024-10-16 17:08:14 +08:00
Yishuo Wang
bb247e991b
refactor merge_qkv and attention_softmax ( #12213 )
2024-10-16 15:58:14 +08:00
Yishuo Wang
e279148aa0
optimize llama3.2 vision again ( #12211 )
2024-10-16 14:29:48 +08:00
Yishuo Wang
f6611f9d3a
optimize llama3.2 vison attention again ( #12204 )
2024-10-15 16:08:20 +08:00
Yishuo Wang
9b81236a2e
optimzie qwen2-vl vision ( #12203 )
2024-10-15 15:54:25 +08:00
Yishuo Wang
d5344587ab
optimize internvl2 vision model's attention ( #12198 )
2024-10-15 10:51:00 +08:00
Yuwen Hu
f8d1adc573
Fix Llama 3.2 & 3.1 on LNL ( #12196 )
2024-10-14 17:39:20 +08:00
Ruonan Wang
310f18c8af
update NPU pipeline generate ( #12182 )
...
* update
* fix style
2024-10-11 17:39:20 +08:00
Ruonan Wang
4d93bb81fe
Initial support of NPU level0 Model ( #12177 )
...
* first commit to support load dll and init llm pipeline
* add init generate
* fix style
* small updates
* fix style and check tokens number
2024-10-11 09:45:53 +08:00
Yishuo Wang
535bee5381
fix qwen2 vl again ( #12174 )
2024-10-10 13:50:01 +08:00
Yishuo Wang
78d253165d
optimize qwen2 vl perf again ( #12167 )
2024-10-09 16:43:48 +08:00
Yishuo Wang
644af2a76e
add basic llama 3.2 vision support ( #12163 )
2024-10-08 10:46:48 +08:00
Yishuo Wang
669ff1a97b
fix sd1.5 ( #12129 )
2024-09-26 17:15:16 +08:00
Yishuo Wang
a266528719
optimize llama 3.2 rope ( #12128 )
2024-09-26 16:08:10 +08:00
Yishuo Wang
584c3489e7
add basic support for llama3.2 ( #12125 )
2024-09-26 15:46:19 +08:00
Yishuo Wang
66f419f8b7
fix qwen2 vl ( #12126 )
2024-09-26 15:44:02 +08:00
Yishuo Wang
77af9bc5fa
support passing None to low_bit in optimize_model ( #12121 )
2024-09-26 11:09:35 +08:00
Yishuo Wang
47e0b83cbf
optimize sd 1.5 ( #12119 )
2024-09-25 15:45:13 +08:00
Yishuo Wang
5d63aef60b
optimize qwen2 vl again ( #12109 )
2024-09-23 13:22:01 +08:00
Ruonan Wang
03bd01c99c
optimize npu qwen2 ( #12107 )
2024-09-20 19:46:16 +08:00
Yishuo Wang
9239fd4f12
add basic support and optimization for qwen2-vl ( #12104 )
2024-09-20 17:23:06 +08:00
Yuwen Hu
828fa01ad3
[NPU] Add mixed_precision for Qwen2 7B ( #12098 )
...
* Add mix_precision argument to control whether use INT8 lm_head for Qwen2-7B-Instruct
* Small fix
* Fixed on load low bit with mixed precision
* Small fix
* Update example accordingly
* Update for default prompt
* Update base on comments
* Final fix
2024-09-20 16:36:21 +08:00
Ruonan Wang
09b8c80d9d
update code for NPU qwen2 ( #12094 )
...
* update code
* fix
2024-09-20 15:58:32 +08:00
Yuwen Hu
f7fb3c896c
Update lm_head optimization for Qwen2 7B ( #12090 )
2024-09-18 17:02:02 +08:00
Wang, Jian4
40e463c66b
Enable vllm load gptq model ( #12083 )
...
* enable vllm load gptq model
* update
* update
* update
* update style
2024-09-18 14:41:00 +08:00
Ruonan Wang
081af41def
[NPU] Optimize Qwen2 lm_head to use INT4 ( #12072 )
...
* temp save
* update
* fix
* fix
* Split lm_head into 7 parts & remove int8 for lm_head when sym_int4
* Simlify and add condition to code
* Small fix
* refactor some code
* fix style
* fix style
* fix style
* fix
* fix
* temp sav e
* refactor
* fix style
* further refactor
* simplify code
* meet code review
* fix style
---------
Co-authored-by: Yuwen Hu <yuwen.hu@intel.com>
2024-09-14 15:26:46 +08:00
Ch1y0q
b4b8c3e495
add lowbit_path for generate.py, fix npu_model ( #12077 )
...
* add `lowbit_path` for `generate.py`, fix `npu_model`
* update `README.md`
2024-09-13 17:28:05 +08:00
Jinhe
4ca330da15
Fix NPU load error message and add minicpm npu lowbit feat ( #12064 )
...
* fix npu_model raise sym_int4 error
* add load_lowbit
* remove print&perf
2024-09-11 16:56:35 +08:00
Ruonan Wang
a0c73c26d8
clean NPU code ( #12060 )
...
* clean code
* remove time.perf_counter()
2024-09-11 15:10:35 +08:00
Yishuo Wang
d8c044e79d
optimize minicpm3 kv cache ( #12052 )
2024-09-10 16:51:21 +08:00
Guancheng Fu
69c8d36f16
Switching from vLLM v0.3.3 to vLLM 0.5.4 ( #12042 )
...
* Enable single card sync engine
* enable ipex-llm optimizations for vllm
* enable optimizations for lm_head
* Fix chatglm multi-reference problem
* Remove duplicate layer
* LLM: Update vLLM to v0.5.4 (#11746 )
* Enable single card sync engine
* enable ipex-llm optimizations for vllm
* enable optimizations for lm_head
* Fix chatglm multi-reference problem
* update 0.5.4 api_server
* add dockerfile
* fix
* fix
* refine
* fix
---------
Co-authored-by: gc-fu <guancheng.fu@intel.com>
* Add vllm-0.5.4 Dockerfile (#11838 )
* Update BIGDL_LLM_SDP_IGNORE_MASK in start-vllm-service.sh (#11957 )
* Fix vLLM not convert issues (#11817 ) (#11918 )
* Fix not convert issues
* refine
Co-authored-by: Guancheng Fu <110874468+gc-fu@users.noreply.github.com>
* Fix glm4-9b-chat nan error on vllm 0.5.4 (#11969 )
* init
* update mlp forward
* fix minicpm error in vllm 0.5.4
* fix dependabot alerts (#12008 )
* Update 0.5.4 dockerfile (#12021 )
* Add vllm awq loading logic (#11987 )
* [ADD] Add vllm awq loading logic
* [FIX] fix the module.linear_method path
* [FIX] fix quant_config path error
* Enable Qwen padding mlp to 256 to support batch_forward (#12030 )
* Enable padding mlp
* padding to 256
* update style
* Install 27191 runtime in 0.5.4 docker image (#12040 )
* fix rebase error
* fix rebase error
* vLLM: format for 0.5.4 rebase (#12043 )
* format
* Update model_convert.py
* Fix serving docker related modifications (#12046 )
* Fix undesired modifications (#12048 )
* fix
* Refine offline_inference arguments
---------
Co-authored-by: Xiangyu Tian <109123695+xiangyuT@users.noreply.github.com>
Co-authored-by: Jun Wang <thoughts.times@gmail.com>
Co-authored-by: Wang, Jian4 <61138589+hzjane@users.noreply.github.com>
Co-authored-by: liu-shaojun <johnssalyn@outlook.com>
Co-authored-by: Shaojun Liu <61072813+liu-shaojun@users.noreply.github.com>
2024-09-10 15:37:43 +08:00
Ruonan Wang
dc4af02b2a
Fix qwen2 1.5B NPU load error ( #12049 )
2024-09-10 14:41:18 +08:00
Yishuo Wang
abc370728c
optimize minicpm3 again ( #12047 )
2024-09-10 14:19:57 +08:00
Ch1y0q
f0061a9916
remove local import os to fix Baichuan NPU load issue ( #12044 )
2024-09-10 14:13:24 +08:00
Ruonan Wang
640998edea
update inter_pp of qwen2 ( #12041 )
2024-09-10 10:34:17 +08:00
Yishuo Wang
048b4590aa
add basic minicpm3 optimization ( #12039 )
2024-09-09 17:25:08 +08:00
Yishuo Wang
6cedb601e4
remove some useless code ( #12035 )
2024-09-06 17:51:08 +08:00
binbin Deng
d2e1b9aaff
Add input padding during prefill for qwen2-7b ( #12033 )
2024-09-06 16:39:59 +08:00
Ruonan Wang
0d04531ae0
update NPU readme of Qwen2 ( #12032 )
...
* update readme
* update broadcast
2024-09-06 15:02:39 +08:00
Yang Wang
58555bd9de
Optimize broadcast for npu llama ( #12028 )
2024-09-06 13:28:20 +08:00
binbin Deng
845e5dc89e
Support lm_head of minicpm-2b on NPU ( #12019 )
2024-09-05 16:19:22 +08:00
Guoqiong Song
8803242f5c
fix llama on cpu ( #12018 )
2024-09-04 19:17:54 -07:00
Ruonan Wang
9eaff5e47d
add save & load support for NPU optimized model ( #11999 )
...
* add save & load support
* fix style
2024-09-03 20:53:22 +08:00
Yuwen Hu
6eb55653ba
Performance mode strategy update for input_embeds input ( #11997 )
2024-09-03 17:46:16 +08:00
binbin Deng
01099f08ee
Revert prefill logic of qwen2-7b ( #11992 )
2024-09-03 14:45:01 +08:00
Yuwen Hu
659d15defc
Fix wrong attention mask and garbage output for inputs_embeds inputs during lookup generation ( #11989 )
...
* Fix garbage output for input_embeds inputs during lookup generation
* Fix on sliding windows
* Simplify code
2024-09-02 19:09:12 +08:00
binbin Deng
2f3d1bd0ec
hotfix qwen2-7b weight setting ( #11991 )
2024-09-02 18:11:08 +08:00
binbin Deng
a40ea7038d
Fix AttributeError of qwen2-1.5B ( #11990 )
2024-09-02 17:55:10 +08:00
Yang Wang
c48817bd43
Support Qwen2-7b MLP in int4 and transpose_value_cache=True ( #11968 )
2024-09-02 14:37:44 +08:00
Ruonan Wang
573c20bae6
fix npu lm_head cpu condition ( #11976 )
...
* fix
* fix
* fix
* fix stype
* fix style
* fix style
2024-08-30 17:11:26 +08:00
Ruonan Wang
60aa1a2c0f
Initial NPU support for MiniCPM-V-2_6 ( #11966 )
...
* initial pr
* update npu model
* fix
* fix kv cache type
* fix
* small fix
* fix style
* fix model id
* change inter_pp=4
* address comment
* fix
* fix style
* fix
* rebase
2024-08-30 16:34:35 +08:00
SONG Ge
158289d205
[NPU] Add initial support for minicpm-llama-v2.5 ( #11962 )
...
* add initial support for minicpm-llama-v2.5
* update impl
* add minicpm-llama3-v2.5 example
2024-08-30 16:00:33 +08:00
binbin Deng
cd077881f1
Disable lm head ( #11972 )
2024-08-30 11:05:18 +08:00
Yang Wang
fbf088f61e
remove obselete npu code ( #11967 )
2024-08-29 14:16:44 -07:00
Yuwen Hu
a9e485eb1b
Support MiniCPM-V-2_6 multi-modal benchmarking with latency text streamer ( #11963 )
...
* Support MiniCPM-V-2_6 multi-modal benchmarking with latency text streamer
* Style fixes
2024-08-29 19:22:09 +08:00
Yina Chen
882f4a5ff7
Add lnl npu driver recommend version and enable cpu_lm_head on llama3 ( #11952 )
...
* update lnl npu driver version and enable cpu_lm_head on llama3
* update
* fix style
* typo
* address comments
* update
* add qwen2-7b
2024-08-29 15:01:18 +08:00
