riaz.somc/transformers
1
0
Fork 0
mirror of https://github.com/huggingface/transformers.git synced 2025-07-31 02:02:21 +06:00
Code Issues Actions 33 Packages Projects Releases Wiki Activity
17,674 Commits 887 Branches 232 Tags 2.1 GiB
e5fd865eba
Commit Graph

17674 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Yijun Lee
e5fd865eba
Add Gemma2 GGUF support (#34002) 
* initial setup for ggml.py

* initial setup of GGUFGemma2Converter class

* Add gemma2 model to gguf.md doc

* Partial work on GGUF_TENSOR_MAPPING

* initial setup of GGUF_TENSOR_MAPPING for Gemma2

* refactor: rename GemmaConvert class to GemmaConverter for naming consistency

* feat: complete gemma2 tensor mapping implementation

* feat: add initial implementation of GGUFGemmaConverter

* feat: complete GGUFGemmaConverter implementation

* feat: add test code for gemma2

* refactor: minor code cleanup

* refactor: minor code cleanup

* fix: resolve suggestions

* Update tests/quantization/ggml/test_ggml.py

Co-authored-by: Isotr0py <2037008807@qq.com>

---------

Co-authored-by: Isotr0py <2037008807@qq.com>
 2025-01-03 14:50:07 +01:00
湛露先生
1fe2d53d4e
Reuse "if not" logic in image_processing. (#35405) 2025-01-03 14:44:57 +01:00
Jacky Lee
30a9971632
Use sdpa_kernel in tests (#35472) 
* update: use sdpa_kernel

* update: rerun test
 2025-01-03 14:39:52 +01:00
Blanchon
cba49cb2a6
Change is_soundfile_availble to is_soundfile_available (#35030) 2025-01-03 14:37:42 +01:00
hoshi-hiyouga
42865860ec
Fix paligemma warning message (#35486) 
fix log input
 2025-01-02 11:36:53 +01:00
湛露先生
b2b04e86e7
Fix docs typos. (#35465) 
Signed-off-by: zhanluxianshen <zhanluxianshen@163.com>
 2025-01-02 11:29:46 +01:00
Matthew Douglas
6b1e86fd4d
Fix new BNB test failures (#35345) 2025-01-02 11:24:52 +01:00
Tom Aarsen
5b516b06c8
Reintroduce Python 3.9 support for ModernBERT (#35458) 
Co-authored-by: Koichi Yasuoka <yasuoka@kanji.zinbun.kyoto-u.ac.jp>
 2025-01-02 11:23:07 +01:00
Jacky Lee
919220dab1
Update translated docs for sdpa_kernel (#35461) 
* docs: update sdpa_kernel for translation

* fix: nn.attention

* update: infer many
 2024-12-31 08:37:58 -08:00
Ahmed Almaghz
eb2b452432
[i18n-ar] Translated file: docs/source/ar/tasks/summarization.md into Arabic (#35195) 
* إضافة الترجمة العربية: summarization.md

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/summarization.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update _toctree.yml

---------

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>
 2024-12-31 08:35:54 -08:00
Ahmed Almaghz
d5aebc6465
[i18n-ar] Translated file: docs/source/ar/tasks/question_answering.md into Arabic (#35196) 
* إضافة الترجمة العربية: question_answering.md

* Update question_answering.md

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update docs/source/ar/tasks/question_answering.md

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>

* Update _toctree.yml

---------

Co-authored-by: Abdullah Mohammed <554032+abodacs@users.noreply.github.com>
 2024-12-30 11:56:05 -08:00
Jacky Lee
b5f97977ed
Update docs for sdpa_kernel (#35410) 
update: sdp_kernel -> sdpa_kernel
 2024-12-30 09:50:34 -08:00
Cheng-Han Chiang
5cabc75b4b
Add compute_loss_func to Seq2SeqTrainer (#35136) 2024-12-29 15:01:35 +01:00
Martin
90f256c90c
Update perf_infer_gpu_one.md: fix a typo (#35441) 2024-12-29 14:57:08 +01:00
Pavel Iakubovskii
5c75087aee
Fix model_accepts_loss_kwargs for timm model (#35257) 
* Fix for timm model

* Add comment
 2024-12-27 16:33:44 +00:00
Kyle Safran
3b0a94ef9e
Fix f-string to show ACCELERATE_MIN_VERSION on error (#35189) 
fix f-string to show ACCELERATE_MIN_VERSION on error

Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>
 2024-12-27 13:21:44 +01:00
Thien Tran
f63da20a9f
CLIP conversion script - Change fairseq to OpenAI (#35384) 
Change fairseq to OpenAI
 2024-12-27 13:12:32 +01:00
宁宇
7f97d01675
Fix: Rename keyword argument in_channels to num_channels (#35289) 
Fix: Rename keyword argument in_channels to num_channels in some default backbone configs
 2024-12-27 13:07:31 +01:00
Quentin Gallouédec
4eb17b26e7
Drop inplace operation for loss computation with gradient accumulation (#35416) 
Fix inplace loss computation
 2024-12-26 14:58:53 +01:00
Anton Vlasjuk
24c91f095f
[GPTQ, CompressedTensors] Fix unsafe imports and metada check (#34815) 
* fix gptq creation when optimum is not installed + fix metadata checking

* fix compressed tensors as well

* style

* pray for ci luck on flaky tests :prayge:

* trigger ci

---------

Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>
Co-authored-by: Mohamed Mekkouri <93391238+MekkCyber@users.noreply.github.com>
 2024-12-24 19:32:44 +01:00
NielsRogge
6e0515e99c
Add DINOv2 with registers (#35348) 
* added changes from 32905

* fixed mistakes caused by select all paste

* rename diff_dinov2...

* ran tests

* Fix modular

* Fix tests

* Use new init

* Simplify drop path

* Convert all checkpoints

* Add figure and summary

* Update paths

* Update docs

* Update docs

* Update toctree

* Update docs

---------

Co-authored-by: BernardZach <bernardzach00@gmail.com>
Co-authored-by: Zach Bernard <132859071+BernardZach@users.noreply.github.com>
 2024-12-24 13:21:59 +01:00
jiqing-feng
d8c1db2f56
enable non-cuda awq model support without modify version (#35334) 
Signed-off-by: jiqing-feng <jiqing.feng@intel.com>
 2024-12-24 12:36:00 +01:00
Yih-Dar
ccc4a5a59b
Disable .github/workflows/self-comment-ci.yml for now (#35366) 
* disable

* disable

---------

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2024-12-24 10:53:57 +01:00
Yoni Gozlan
93aafdc620
Add compile test for fast image processor (#35184) 
* add compile test for fast image processor

* override pixtral test
 2024-12-23 13:12:45 -05:00
Mohamed Mekkouri
82fcac0a7e
Adding logger.info about update_torch_dtype in some quantizers (#35046) 
adding logger.info
 2024-12-23 17:01:00 +01:00
Miquel Farré
a1780b7ba5
bugfix Idefics3 processor - handle gracefully cases with text and no images (#35363) 
* bugfix processing empty images

* fix

* fix

* Update src/transformers/models/idefics3/processing_idefics3.py

Co-authored-by: Yoni Gozlan <74535834+yonigozlan@users.noreply.github.com>

* adding tests

* fix

* fix

* fix

---------

Co-authored-by: Yoni Gozlan <74535834+yonigozlan@users.noreply.github.com>
 2024-12-23 16:59:01 +01:00
Andrei Panferov
64c05eecd6
HIGGS Quantization Support (#34997) 
* higgs init

* working with crunches

* per-model workspaces

* style

* style 2

* tests and style

* higgs tests passing

* protecting torch import

* removed torch.Tensor type annotations

* torch.nn.Module inheritance fix maybe

* hide inputs inside quantizer calls

* style structure something

* Update src/transformers/quantizers/quantizer_higgs.py

Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>

* reworked num_sms

* Update src/transformers/integrations/higgs.py

Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>

* revamped device checks

* docstring upd

* Update src/transformers/quantizers/quantizer_higgs.py

Co-authored-by: Mohamed Mekkouri <93391238+MekkCyber@users.noreply.github.com>

* edited tests and device map assertions

* minor edits

* updated flute cuda version in docker

* Added p=1 and 2,3bit HIGGS

* flute version check update

* incorporated `modules_to_not_convert`

* less hardcoding

* Fixed comment

* Added docs

* Fixed gemma support

* example in docs

* fixed torch_dtype for HIGGS

* Update docs/source/en/quantization/higgs.md

Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>

* Collection link

* dequantize interface

* newer flute version, torch.compile support

* unittest message fix

* docs update compile

* isort

* ValueError instead of assert

---------

Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>
Co-authored-by: Mohamed Mekkouri <93391238+MekkCyber@users.noreply.github.com>
 2024-12-23 16:54:49 +01:00
Huazhong Ji
ef1f54a0a7
add bnb support for Ascend NPU (#31512) 
* add bnb support for Ascend NPU

* delete comment
 2024-12-23 16:36:16 +01:00
Mohamed Mekkouri
59178780a6
Fix : VPTQ test (#35394) 
fix_test
 2024-12-23 16:27:46 +01:00
Alvaro Bartolome
3a4ced9ab4
Fix typing in docstring for PaliGemmaProcessor (#35278) 
Updated typing for `tokenizer` in the `PaliGemmaProcessor` to be `GemmaTokenizerFast` instead of `LlamaTokenizerFast`
 2024-12-23 16:22:04 +01:00
Quentin Gallouédec
3cd3cd50ac
Scale loss before backward (#35207) 2024-12-23 16:16:38 +01:00
Mohamed Mekkouri
f5264a86ee
Deprecate _is_quantized_training_enabled (#34991) 
deperecate

Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>
 2024-12-23 15:51:31 +01:00
Tibor Reiss
e10be82b71
uniformize kwargs for SAM (#34578) 
* Make kwargs uniform for SAM

* Remove unused attribute

* Make point_pad_value part of image_kwargs

* Update annotations

* Code review - use existing methods

* Use ProcessorTesterMixin

* Do not add ProcessorTesterMixin everywhere
 2024-12-23 13:54:57 +01:00
Taha Yassine
2bb60982ac
Patch GPTNeoX to use adequate FA2 if position_ids is provided (#35318) 2024-12-23 13:45:55 +01:00
Wing Lian
5e7aedebeb
make LlamaModel._update_causal_mask torch compilable (#35187) 
* make LlamaModel._update_causal_mask torch compilable

* chore: lint (make fix-copies)

* fix-copies

---------

Co-authored-by: Arthur Zucker <arthur.zucker@gmail.com>
 2024-12-23 13:10:00 +01:00
Matthew Douglas
401aa39d7b
bitsandbytes: simplify 8bit dequantization (#35068) 2024-12-23 13:04:59 +01:00
Cyril Vallez
05260a1fc1
Fix new FA2 if is_causal is passed explicitly (#35390) 
* fix

* Update modeling_decision_transformer.py

* Update flash_attention.py
 2024-12-22 20:00:07 +01:00
bastrob
8f38f58f3d
owlvit/2 dynamic input resolution (#34764) 
* owlvit/2 dynamic input resolution.

* adapt box grid to patch_dim_h patch_dim_w

* fix ci

* clarify variable naming

* clarify variable naming..

* compute box_bias dynamically inside box_predictor

* change style part of code

* [run-slow] owlvit, owlv2
 2024-12-21 08:51:09 +00:00
Steven Liu
608e163b52
[docs] Follow up register_pipeline (#35310) 
example json
 2024-12-20 09:22:44 -08:00
UV
94fe0b915b
Improved Documentation Of Audio Classification (#35368) 
* Improved Documentation Of Audio Classification

* Updated documentation as per review

* Updated audio_classification.md

* Update audio_classification.md
 2024-12-20 09:17:28 -08:00
Joel Koch
c96cc039c3
Improve modular transformers documentation (#35322) 
* Improve modular transformers documentation

- Adds hints to general contribution guides
- Lists which utils scripts are available to generate single-files from modular files and check their content

* Show commands in copyable code cells

---------

Co-authored-by: Joel Koch <joel@bitcrowd.net>
 2024-12-20 09:16:02 -08:00
Yih-Dar
504c4d3692
Make test_generate_with_static_cache even less flaky (#34995) 
* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

---------

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2024-12-20 16:03:26 +01:00
Yih-Dar
0fc2970363
Use weights_only=True with torch.load for transfo_xl (#35241) 
fix

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2024-12-20 15:40:55 +01:00
Arthur
6fae2a84ae
Update test fetcher when we want to test all (#35364) 
* [test-all]

* style

* [test-all]

* [test_all]

* [test_all]

* style
 2024-12-20 15:10:43 +01:00
nhamanasu
34ad1bd287
update codecarbon (#35243) 
* update codecarbon

* replace directly-specified-test-dirs with tmp_dir

* Revert "replace directly-specified-test-dirs with tmp_dir"

This reverts commit 310a6d962e.

* revert the change of .gitignore

* Update .gitignore

---------

Co-authored-by: Yih-Dar <2521628+ydshieh@users.noreply.github.com>
 2024-12-20 15:04:36 +01:00
Jiwoong
40292aa4e9
bugfix: torch.export failure caused by _make_causal_mask (#35291) 
* bugfix: torch.export failure caused by `_make_causal_mask`

Recent changes in torch dynamo prevent mutations on tensors converted with aten::_to_copy. To address this, we can clone such tensor before performing in-place operation `masked_fill_` only when the code is being compiled by torch dynamo.
(relevant issue: https://github.com/pytorch/pytorch/issues/127571)

* chore: use `is_torchdynamo_compiling` instead of `torch._dynamo.is_compiling`
 2024-12-20 14:37:04 +01:00
Yih-Dar
05de764e9c
Aurevoir PyTorch 1 (#35358) 
* fix

* fix

* fix

---------

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2024-12-20 14:36:31 +01:00
Qizhi Chen
4567ee8057
fix zoedepth initialization error under deepspeed zero3 (#35011) 
fix zoe bug in deepspeed zero3
 2024-12-20 11:42:40 +00:00
Jacky Lee
c3a43594b7
Add Tensor Parallel support for Qwen2VL (#35050) 
feat: add parallel support for qwen2vl
 2024-12-20 12:40:38 +01:00
Cyril Vallez
0d51d65905
Cleaner attention interfaces (#35342) 
* cleaner attention interfaces

* correctly set the _attn_implementation when adding other functions to it

* update

* Update modeling_utils.py

* CIs
 2024-12-20 12:09:34 +01:00