Code

ipex 2.3 released#725

Code

Merged

echarlaix merged 50 commits intohuggingface:mainhuggingface/optimum-intel:mainfrom jiqing-feng:renamejiqing-feng/optimum-intel:renameCopy head branch name to clipboardJun 6, 2024

+164-56

Commits

Commits on May 23, 2024

• ipex 2.3 released

  jiqing-feng

  committed
  5351f4a

  Copy full SHA for 5351f4a

Commits on May 27, 2024

• skip tests

  jiqing-feng

  committed
  1f98d6d

  Copy full SHA for 1f98d6d

• skip testing without pkv

  jiqing-feng

  committed
  b2b93bb

  Copy full SHA for b2b93bb

• add tests skip

  jiqing-feng

  committed
  64dcde4

  Copy full SHA for 64dcde4

• only llama2 with at least 64 head size support IAKV

  jiqing-feng

  committed
  945f6b6

  Copy full SHA for 945f6b6

• cannot assert same outputs cause do_sample=True

  jiqing-feng

  committed
  c8922f3

  Copy full SHA for c8922f3

• rm tiny-llama model testing cause it not work for IAKV

  jiqing-feng

  committed
  2ddfa7a

  Copy full SHA for 2ddfa7a

Commits on May 28, 2024

• fix code style

  jiqing-feng

  committed
  f4e887d

  Copy full SHA for f4e887d

• fix style

  jiqing-feng

  committed
  d96ea58

  Copy full SHA for d96ea58

• rm tiny llama on test pipeline

  jiqing-feng

  committed
  ec24d5a

  Copy full SHA for ec24d5a

Commits on May 30, 2024

• fix tests

  jiqing-feng

  committed
  871de7b

  Copy full SHA for 871de7b

• support use_cache=False

  jiqing-feng

  committed
  d0c8951

  Copy full SHA for d0c8951

• rm use_cache in model_kwargs

  jiqing-feng

  committed
  537f0aa

  Copy full SHA for 537f0aa

• set use_cache

  jiqing-feng

  committed
  5a71790

  Copy full SHA for 5a71790

Commits on May 31, 2024

• Update optimum/intel/ipex/modeling_base.py

  Show description for bde814e
  

  jiqing-feng

  and

  echarlaix

  authored
  bde814e

  Copy full SHA for bde814e

• fix spelling error

  jiqing-feng

  committed
  4a81ea9

  Copy full SHA for 4a81ea9

• fix style

  jiqing-feng

  committed
  3a61e84

  Copy full SHA for 3a61e84

• add transformers version warning

  jiqing-feng

  committed
  fd69407

  Copy full SHA for fd69407

• add compare resultes

  jiqing-feng

  committed
  1032a26

  Copy full SHA for 1032a26

• add warning

  jiqing-feng

  committed
  c8e7969

  Copy full SHA for c8e7969

• set pad_token_id

  jiqing-feng

  committed
  afdc8d7

  Copy full SHA for afdc8d7

Commits on Jun 3, 2024

• limited transformers

  jiqing-feng

  committed
  1d1df34

  Copy full SHA for 1d1df34

• fix transformers version

  jiqing-feng

  committed
  aaaa4c3

  Copy full SHA for aaaa4c3

Commits on Jun 4, 2024

• update transformers version

  jiqing-feng

  committed
  f6b8010

  Copy full SHA for f6b8010

• fix version

  jiqing-feng

  committed
  51e47b6

  Copy full SHA for 51e47b6

• temporary fix for multi-query model

  jiqing-feng

  committed
  5204b24

  Copy full SHA for 5204b24

• fix code styke

  jiqing-feng

  committed
  8f2f025

  Copy full SHA for 8f2f025

• add transformers version tests

  jiqing-feng

  committed
  8dc5ad5

  Copy full SHA for 8dc5ad5

Commits on Jun 5, 2024

• Update .github/workflows/test_ipex.yml

  Show description for e482e58
  

  jiqing-feng

  and

  echarlaix

  authored
  e482e58

  Copy full SHA for e482e58

• check geenration method

  jiqing-feng

  committed
  d366b80

  Copy full SHA for d366b80

• Update optimum/intel/ipex/modeling_base.py

  Show description for 3948cad
  

  jiqing-feng

  and

  echarlaix

  authored
  3948cad

  Copy full SHA for 3948cad

• fix use_cache

  jiqing-feng

  committed
  d1b63ef

  Copy full SHA for d1b63ef

• add hidden size limitation for patch

  jiqing-feng

  committed
  ea4d3e2

  Copy full SHA for ea4d3e2

• add llama in tests

  jiqing-feng

  committed
  bcb2b5a

  Copy full SHA for bcb2b5a

• add re-load tests

  jiqing-feng

  committed
  f5f1af8

  Copy full SHA for f5f1af8

• fix hidden size check

  jiqing-feng

  committed
  c08c957

  Copy full SHA for c08c957

• rm norm config

  jiqing-feng

  committed
  51e6f3d

  Copy full SHA for 51e6f3d

• add version variable

  jiqing-feng

  committed
  d06123b

  Copy full SHA for d06123b

• fix import

  jiqing-feng

  committed
  641e8f9

  Copy full SHA for 641e8f9

• rm useless logger

  jiqing-feng

  committed
  50c1059

  Copy full SHA for 50c1059

• rm useless logging

  jiqing-feng

  committed
  a961746

  Copy full SHA for a961746

Commits on Jun 6, 2024

• fix last round review

  jiqing-feng

  committed
  c2253a8

  Copy full SHA for c2253a8

• Merge branch 'huggingface:main' into rename

  jiqing-feng

  authored
  e29ea58

  Copy full SHA for e29ea58

• Update .github/workflows/test_ipex.yml

  echarlaix

  authored
  caa27c3

  Copy full SHA for caa27c3

• Update optimum/intel/ipex/modeling_base.py

  echarlaix

  authored
  78498ab

  Copy full SHA for 78498ab

• Update optimum/intel/ipex/modeling_base.py

  echarlaix

  authored
  97f7876

  Copy full SHA for 97f7876

• Update setup.py

  echarlaix

  authored
  cf3525a

  Copy full SHA for cf3525a

• Update optimum/exporters/ipex/modeling_utils.py

  echarlaix

  authored
  8ba602d

  Copy full SHA for 8ba602d

• fix

  jiqing-feng

  committed
  f15a1f5

  Copy full SHA for f15a1f5

• limit the new tokens of assisted decoding tests

  jiqing-feng

  committed
  36ae751

  Copy full SHA for 36ae751