saymrwulf/transformers
1
0
Fork 0
mirror of https://github.com/saymrwulf/transformers.git synced 2026-05-14 20:58:08 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Commit graph

  • 737ee3af62
         Merge branch 'main' into feature/#35425 Minho Ryu 2025-01-28 14:46:31 +0900
  • 704767e05c add deepseekv3 modeling ryan u 2025-01-28 14:42:30 +0900
  • a16e46b080 Use process_retry on amd-smi Ivar Flakstad 2025-01-27 14:40:36 +0100
  • 4d15ba4458
         Merge branch 'main' into remove-torch-pre-releases-amd-image remove-torch-pre-releases-amd-image ivarflakstad 2025-01-27 14:32:46 +0100
  • 37df1312fc Do not use pre-releases for torch libs Ivar Flakstad 2025-01-27 13:18:23 +0100
  • 9d4657c3be fix osme missing atols fix-red-ci-atol Arthur Zucker 2025-01-27 11:04:34 +0100
  • a33dd6e488
         Merge branch 'main' into make-cache-traceable make-cache-traceable Ilyas Moutawwakil 2025-01-26 20:40:50 +0100
  • 5a2ff5dfb0 Set to pull_request_target, testing works! auto-assign-reviewers Matt 2025-01-24 15:42:53 +0000
  • fdaacaaf03 Set back to pull_request for testing Matt 2025-01-24 15:40:34 +0000
  • feafbe087e Update the script Matt 2025-01-24 15:40:15 +0000
  • f124ec012b Use pull-request-target instead Matt 2025-01-24 14:57:26 +0000
  • 090d9c4b2a
         Merge branch 'main' into tensor-cache tensor-cache Ilyas Moutawwakil 2025-01-24 12:02:45 +0100
  • e13e9c1e25 fix gemma that needed kwargs fix-kwargs-issues Arthur Zucker 2025-01-24 11:19:47 +0100
  • 67dd5524d3 simply make cache traceable IlyasMoutawwakil 2025-01-24 11:19:36 +0100
  • 016ae273a2 Add TODO Matt 2025-01-23 17:50:00 +0000
  • 5ccb79c16d fixed dynamic cache IlyasMoutawwakil 2025-01-23 16:45:28 +0100
  • 2d480eccc7 fix copies tp-support Arthur Zucker 2025-01-23 16:40:33 +0100
  • 3a12f71ab9 Request reviews instead of assigning Matt 2025-01-22 20:10:49 +0000
  • 580aa713cf Request reviews instead of assigning Matt 2025-01-22 20:05:56 +0000
  • 8b20315634 Remove prefix Matt 2025-01-22 19:53:16 +0000
  • adad02848a Strip inline comments Matt 2025-01-22 19:39:14 +0000
  • 27d2961545 Update debug logs Matt 2025-01-22 19:35:34 +0000
  • 3d6105a8d8 Update workflow permissions Matt 2025-01-22 19:29:17 +0000
  • 8dc084682c Update workflow permissions Matt 2025-01-22 19:23:27 +0000
  • 6b0f5b9b24 Correct path for codeowners file Matt 2025-01-22 19:15:51 +0000
  • ef3df762f3 Temporarily comment out the opened line so we can test the script Matt 2025-01-22 19:13:07 +0000
  • e96ba83ad4 Don't reassign reviewers if we already have them Matt 2025-01-22 19:08:37 +0000
  • 4333c61971 fix missing import Matt 2025-01-22 19:06:54 +0000
  • e17ab9831e First draft of github action on PR opening for auto-assigning reviewers Matt 2025-01-22 19:04:36 +0000
  • 80b49d721b rebased IlyasMoutawwakil 2025-01-22 17:31:39 +0100
  • dc1bd15ba9 Merge branch 'main' into tensor-cache IlyasMoutawwakil 2025-01-22 17:30:23 +0100
  • 338f5954b9 more reverts IlyasMoutawwakil 2025-01-22 17:29:48 +0100
  • 2f4e0bc93e
         Update src/transformers/cache_utils.py Ilyas Moutawwakil 2025-01-22 17:18:28 +0100
  • 485f959f85 revert IlyasMoutawwakil 2025-01-22 17:17:17 +0100
  • 2bbbbbcf97 add device and dtype setters IlyasMoutawwakil 2025-01-22 17:15:12 +0100
  • 85c71b004b
         Merge branch 'main' into tensor-cache Ilyas Moutawwakil 2025-01-22 15:53:33 +0100
  • da60604f2c fix test_cache_utils IlyasMoutawwakil 2025-01-22 15:43:14 +0100
  • 6e9799c817 add clone and to IlyasMoutawwakil 2025-01-22 15:42:43 +0100
  • 4950a9e3f0 extract wrapper kwargs from init signature to correctly instantate IlyasMoutawwakil 2025-01-22 13:49:01 +0100
  • 2af7730cb2 More updates to timm image processor, kwarg handling * default to train input size (less surprising) * add properties to mimic .size .crop_size .image_mean .image_std attributes in many Transformers image preproc (works with autotrain now) * try to make key check / inspect code more clear timm_wrapper_kwargs Ross Wightman 2025-01-21 15:57:54 -0800
  • da30662b81 Exploring use of kwargs for timm model and transforms creation Ross Wightman 2025-01-21 08:47:25 -0800
  • f2acf5fe34 Expound x2 muellerzr-trainer-refactor [[ -z $EMAIL ]] && read -e -p "Enter your email (for git configuration): " EMAIL 2025-01-21 09:25:18 -0500
  • 59e10153da Document what's happening in the code [[ -z $EMAIL ]] && read -e -p "Enter your email (for git configuration): " EMAIL 2025-01-21 09:24:17 -0500
  • a0ce95c7dc Readbility muellerzr-more-ga-tests-fast [[ -z $EMAIL ]] && read -e -p "Enter your email (for git configuration): " EMAIL 2025-01-21 09:02:06 -0500
  • 5d1545370e better error message circleci_debug_base_MobileNetV1ModelTest_test_batching_equivalence ydshieh 2025-01-21 12:41:05 +0100
  • 063286f228
         Remove cache migration script remove-cache-migration-script Wauplin 2025-01-21 11:37:39 +0100
  • 57c02ccf15 bump rocm image build_ci_docker_image_amd2 Ivar Flakstad 2025-01-20 20:31:16 +0100
  • c075d2cd62 Fix AutoProcessor import order issue with custom classes fix-autoprocessor-import-order openhands 2025-01-20 18:14:34 +0000
  • b67b6eb9b2 make cache class exportable and executorch compatible IlyasMoutawwakil 2025-01-20 18:47:30 +0100
  • 78257cac9f skip ydshieh 2025-01-20 18:00:35 +0100
  • 1212cb5eae fix ydshieh 2025-01-20 17:31:57 +0100
  • 53e70d9c69 fix ydshieh 2025-01-20 17:26:13 +0100
  • d269417aab fix zamba and jamba dynamic cache IlyasMoutawwakil 2025-01-20 17:21:49 +0100
  • 95c1686ee0 style IlyasMoutawwakil 2025-01-20 17:09:21 +0100
  • 8606594ad4 fix boolean evaluation IlyasMoutawwakil 2025-01-20 17:08:37 +0100
  • 2e752ead46 revert my changes v4.48.1 Arthur Zucker 2025-01-20 17:05:34 +0100
  • 45bb39bb80 torch tensor subclassing IlyasMoutawwakil 2025-01-20 17:01:49 +0100
  • 785b5cf444 v4.48.1 Arthur Zucker 2025-01-20 16:20:06 +0100
  • 3b09464364 Patch moonshine (#35731) eustlb 2025-01-20 16:19:29 +0100
  • b00807fac2 Fix condition when GA loss bug fix is not performed (#35651) kang sheng 2025-01-16 20:59:53 +0800
  • 612bfd0801 [Phi] bias should be True (#35650) Arthur 2025-01-13 13:15:07 +0100
  • a77a94b209 unproxy cache IlyasMoutawwakil 2025-01-20 14:43:41 +0100
  • d4b631edd0 use tensor cache instead of module cache IlyasMoutawwakil 2025-01-20 14:17:28 +0100
  • 8a462d13d3
         Merge branch 'main' into secure-amd-ci secure-amd-ci ivarflakstad 2025-01-17 20:47:18 +0100
  • 4afffcf9a6 Revert some changes that were deemed no longer required Ivar Flakstad 2025-01-17 20:46:17 +0100
  • ef0b5e279c add more TP support Arthur Zucker 2025-01-17 11:27:13 +0100
  • 9f6481796d fix the small freeblocks issue continuous-batching Arthur Zucker 2025-01-16 15:13:44 +0100
  • f56824b0cb update Arthur Zucker 2025-01-16 14:19:25 +0100
  • cdd1d6e44c finish working example Arthur Zucker 2025-01-16 14:13:15 +0100
  • fac571ac65 don't loop too much Arthur Zucker 2025-01-16 11:40:07 +0100
  • aafc48b654 nits and fixes Arthur Zucker 2025-01-16 11:38:17 +0100
  • 74e09dc4e0 works! Arthur Zucker 2025-01-16 11:24:54 +0100
  • 32e7e7b6b1 make style fix_quanto_llama27b MekkCyber 2025-01-15 17:15:49 +0000
  • 76815d1360 fix_quanto MekkCyber 2025-01-15 17:15:38 +0000
  • 517cae97bb up Arthur Zucker 2025-01-15 18:07:25 +0100
  • c800a2c913 up Arthur Zucker 2025-01-15 17:51:12 +0100
  • 960e176910 small updated Arthur Zucker 2025-01-15 17:47:36 +0100
  • 3fc1e02e3c initial commit Arthur Zucker 2025-01-15 16:55:04 +0100
  • 19c73cb0b1 Remove redundant variable Ivar Flakstad 2025-01-15 13:02:21 +0100
  • b0a095ba50 Merge branch 'main' into secure-amd-ci Ivar Flakstad 2025-01-15 12:08:50 +0100
  • 526bb303d2 Fix call to get_workflow_id. ruff format Ivar Flakstad 2025-01-15 12:07:47 +0100
  • cc6f662a54 Testing success, remove debug block faster_set_initialized_submodules Matt 2025-01-14 18:24:57 +0000
  • edda0c1390 Formatting cleanup Matt 2025-01-14 18:12:04 +0000
  • dcbc8c9cce Fix the old keys comparison Matt 2025-01-14 18:05:07 +0000
  • 3ec087ed73 make fixup Matt 2025-01-14 17:49:49 +0000
  • 88aac166db Make set_initialized_submodules O(kN + log(N)) instead of O(N^2), where k << N Matt 2025-01-14 17:42:33 +0000
  • 0d90a51f72 Add workflow_id (defaults to Self-hosted runner (scheduled)) Ivar Flakstad 2025-01-14 15:00:18 +0100
  • da3448dacf handle empty string REPORT_REPO_ID correctly Ivar Flakstad 2025-01-14 14:35:47 +0100
  • 0652d891a7 Actually fix in the modular file fix-gemma2-sliding-window Pedro Cuenca 2025-01-14 12:24:29 +0100
  • 6564e152ed Fix Gemma2 sliding window attention Pedro Cuenca 2025-01-14 12:15:48 +0100
  • 637cadb26b test on transformers-supported revision fix_aria_ci Pablo 2025-01-13 18:23:53 +0100
  • b0be2eda9b Re-add space [[ -z $EMAIL ]] && read -e -p "Enter your email (for git configuration): " EMAIL 2025-01-13 11:27:52 -0500
  • 7306624f45 Further nits [[ -z $EMAIL ]] && read -e -p "Enter your email (for git configuration): " EMAIL 2025-01-13 11:24:12 -0500
  • 776758b597 Add more rigerous non-slow grad accum tests [[ -z $EMAIL ]] && read -e -p "Enter your email (for git configuration): " EMAIL 2025-01-13 11:18:41 -0500
  • b73bf1d1bd [run-slow] bamba fix_bamba_test Pablo 2025-01-13 15:59:52 +0100
  • e2cb0b96d1 make explicit gpu dep Pablo 2025-01-13 15:57:59 +0100
  • e00858ffd6 stash for now Arthur Zucker 2025-01-13 09:41:19 +0100
  • 6bc0fbcfa7 [WIP] Emu3: add model (#33770) v4.48.0 Raushan Turganbay 2025-01-10 12:23:00 +0100
  • 59e28c30fa Fix flex_attention in training mode (#35605) Cyril Vallez 2025-01-10 11:49:12 +0100
  • 7cf6230e25 push a fix for now Arthur Zucker 2025-01-10 11:34:08 +0100