SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity
10,596 Commits 5 Branches 239 Tags
90f6fe9155d2f477588d9ba2d7c697a1933e205a
Commit Graph

10596 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Steven Liu
90f6fe9155 Skip some doctests in quicktour (#18927) 
* skip some code examples for doctests

* make style

* fix code snippet formatting

* separate code snippet into two blocks
 2022-09-07 14:45:22 -07:00
lewtun
6519150c31 Add image height and width to ONNX dynamic axes (#18915) 2022-09-07 22:42:46 +02:00
Colin Dean
737f6ad1f7 Starts on a list of external deps required for dev (#18929) 
* Starts on a list of external deps required for dev

I've found that I need to install MeCab manually on my AS Mac.

* Generalizes OS nascent dependency list

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
 2022-09-07 16:33:03 -04:00
Yanming Wang
6394221871 Fix XLA fp16 and bf16 error checking (#18913) 
* Fix XLA fp16 and bf16 error checking

* Update src/transformers/training_args.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
 2022-09-07 15:45:17 -04:00
Yih-Dar
6690ba3f4d pin TF 2.9.1 for self-hosted CIs (#18925) 
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2022-09-07 19:46:14 +02:00
Ankur Goyal
2ef7742117 Add DocumentQuestionAnswering pipeline (#18414) 
* [WIP] Skeleton of VisualQuestionAnweringPipeline extended to support LayoutLM-like models

* Fixup

* Use the full encoding

* Basic refactoring to DocumentQuestionAnsweringPipeline

* Cleanup

* Improve args, docs, and implement preprocessing

* Integrate OCR

* Refactor question_answering pipeline

* Use refactored QA code in the document qa pipeline

* Fix tests

* Some small cleanups

* Use a string type annotation for Image.Image

* Update encoding with image features

* Wire through the basic docs

* Handle invalid response

* Handle empty word_boxes properly

* Docstring fix

* Integrate Donut model

* Fixup

* Incorporate comments

* Address comments

* Initial incorporation of tests

* Address Comments

* Change assert to ValueError

* Comments

* Wrap `score` in float to make it JSON serializable

* Incorporate AutoModeLForDocumentQuestionAnswering changes

* Fixup

* Rename postprocess function

* Fix auto import

* Applying comments

* Improve docs

* Remove extra assets and add copyright

* Address comments

Co-authored-by: Ankur Goyal <ankur@impira.com>
 2022-09-07 13:38:49 -04:00
Olatunji Ruwase
3059d80d80 [DeepSpeed ZeRO3] Fix performance degradation in sharded models (#18911) 
* [DeepSpeed] Fix performance degradation in sharded models

* style

* polish

Co-authored-by: Stas Bekman <stas@stason.org>
 2022-09-07 07:44:20 -07:00
Yih-Dar
10c774cf60 remvoe _create_and_check_torch_fx_tracing in specific test files (#18667) 
* remvoe _create_and_check_torch_fx_tracing defined in specific model test files

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2022-09-07 16:22:09 +02:00
Joao Gante
0eabab0998 TF: final bias as a layer in seq2seq models (replicate TFMarian fix) (#18903) 2022-09-07 14:03:02 +01:00
Matt
2b9513fdab Update TF fine-tuning docs (#18654) 
* Update TF fine-tuning docs

* Fix formatting

* Add some section headers so the right sidebar works better

* Squiggly it

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/training.mdx

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Explain things in the text, not the comments

* Make the two dataset creation methods into a list

* Move the advice about collation out of a <Tip>

* Edits for clarity

* Edits for clarity

* Edits for clarity

* Replace `to_tf_dataset` with `prepare_tf_dataset` in the fine-tuning pages

* Restructure the page a little bit

* Restructure the page a little bit

* Restructure the page a little bit

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
 2022-09-07 13:30:07 +01:00
Wang, Yi
d842f2d5b9 update the train_batch_size in case HPO change batch_size_per_device (#18918) 
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>

Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
 2022-09-07 08:01:30 -04:00
Nicholas Broad
4f299b2446 Accelerator end training (#18910) 
* add accelerator.end_training()

Some trackers need this to end their runs.

* fixup and quality

* add space

* add space again ?!?
 2022-09-07 07:46:26 -04:00
Yih-Dar
7a8118947f Add checks for more workflow jobs (#18905) 
* add check for scheduled CI

* Add check to other CIs

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2022-09-07 12:51:37 +02:00
NielsRogge
c25f27fa6a [VideoMAE] Improve code examples (#18919) 
* Simplify code example

* Add seed
 2022-09-07 12:24:12 +02:00
Ekagra Ranjan
0a632f076d Fix incorrect size of input for 1st strided window length in Perplexity of fixed-length models (#18906) 
* update the PPL for stride 512

* fix 1st strided window size

* linting

* fix typo

* styling
 2022-09-06 15:20:12 -04:00
Yih-Dar
7d5fde991d unpin slack_sdk version (#18901) 
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2022-09-06 18:42:00 +02:00
Sylvain Gugger
71ff88fa4f Further reduce the number of alls to head for cached objects (#18871) 
* Further reduce the number of alls to head for cached models/tokenizers/pipelines

* Fix tests

* Address review comments
 2022-09-06 12:34:37 -04:00
Alara Dirik
6678350c01 fixes bugs to handle non-dict output (#18897) 2022-09-06 16:13:34 +03:00
Yih-Dar
998a90bc7d Fix test_tf_encode_plus_sent_to_model for LayoutLMv3 (#18898) 
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2022-09-06 14:51:03 +02:00
Ekagra Ranjan
f85acb4d73 Fix decode_input_ids to bare T5Model and improve doc (#18791) 
* use tokenizer to output tensor

* add preprocessing for decoder_input_ids for bare T5Model

* add preprocessing to tf and flax

* linting

* linting

* Update src/transformers/models/t5/modeling_flax_t5.py

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

* Update src/transformers/models/t5/modeling_tf_t5.py

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

* Update src/transformers/models/t5/modeling_t5.py

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
 2022-09-06 14:12:26 +02:00
arun99481
3b19c0317b updating gather function with gather_for_metrics in run_wav2vec2_pretraining (#18877) 
Co-authored-by: Arun Rajaram <arunrajaram@Aruns-MacBook-Pro.local>
 2022-09-06 07:36:37 -04:00
Had
734b7e2a5a Mask t5 relative position bias then head pruned (#17968) 
* add position bias head masking if heads pruned

* fix pruning function in t5 encoder

* make style

* make fix-copies

* Revert added folder

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
 2022-09-06 10:39:31 +02:00
Joao Gante
d4dbd7ca59 Generate: get the correct beam index on eos token (#18851) 2022-09-05 19:35:47 +01:00
zkep
c6d3daba54 Update Chinese documentation (#18893) 
* update the translation
 2022-09-05 19:56:12 +02:00
Sofia Oliveira
cfd623a859 Add type hints to XLM-Roberta-XL models (#18475) 
* Add type hints to XLM-Roberta-XL models

* Format
 2022-09-05 13:38:08 +01:00
Surya Prakash Sahu
17c634fd5b Update perf_train_gpu_one.mdx (#18442) 2022-09-05 14:06:36 +02:00
Patrick von Platen
badb9d2aaa Correct naming pegasus x (#18896) 
* add first generation tutorial

* [Pegasus X] correct naming

* [Generation] Remove
 2022-09-05 11:25:00 +02:00
Lysandre Debut
591cfc6c90 Mention TF and Flax checkpoints (#18894) 2022-09-05 11:09:39 +02:00
Joao Gante
7f27e002fd TF: TFMarianMTModel final logits bias as a layer (#18833) 
* bias as a layer

* alias the bias (hah, it rhymes)

* add comment with info
 2022-09-05 09:20:27 +01:00
Steven Liu
65fb71bc76 Add Trainer to quicktour (#18723) 
* 📝 update quicktour

* 📝 add trainer section

* 🖍 markdown table, apply feedbacks

* ✨ make style

* add tf training section

* make style
 2022-09-02 15:05:31 -05:00
Steven Liu
ae32f3afef Finetune guide for semantic segmentation (#18640) 
* 📝 first draft

* oops add to toctree

* make style

* 📝 add inference section

* 🖍 make style

* 📝 add images

* 🖍 apply feedbacks

* remove num_labels and pytorch block

* apply feedbacks, add colab notebook

Co-authored-by: Steven <stevhliu@gmail.com>
 2022-09-02 14:29:51 -05:00
Steven Liu
bf9d506137 Update docs landing page (#18590) 
* 📝 update docs landing page

* 🖍 apply feedbacks

* apply feedbacks

* apply feedbacks, use <br> for list
 2022-09-02 14:29:06 -05:00
Jason Phang
53e33e6f1b PEGASUS-X (#18551) 
* PegasusX Initial commit

* rename

* pegasus X implementation

* pegx update

* pegx fix

* pegasus-x fixes

* pegx updates

* cleanup

* cleanup

* cleanup

* tests

* stylefixes

* Documentation update

* Model hub fix

* cleanup

* update

* update

* testfix

* Check fix

* tweaks for merging

* style

* style

* updates for pr

* style

* change pegasus-x repo
 2022-09-02 19:54:02 +02:00
Yih-Dar
ecdf9b06bc Remove cached torch_extensions on CI runners (#18868) 
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2022-09-02 18:17:58 +02:00
Yih-Dar
4e29b3f884 A script to download artifacts and perform CI error statistics (#18865) 
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2022-09-02 17:59:26 +02:00
Joao Gante
9196f48b95 Generate: validate model_kwargs on TF (and catch typos in generate arguments) (#18651) 2022-09-02 16:25:26 +01:00
Stas Bekman
c5be7cae59 postpone bnb load until it's needed (#18859) 2022-09-02 08:22:46 -07:00
Sylvain Gugger
9e346f7436 Fix number of examples for iterable datasets in multiprocessing (#18856) 
* Fix number of examples for iterable datasets in multiprocessing

* Add stronger check
 2022-09-02 10:49:39 -04:00
Yih-Dar
0ab465a5d2 pin Slack SDK to 3.18.1 to avoid failing issue (#18869) 
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2022-09-02 16:49:08 +02:00
Sylvain Gugger
38c3cd52fb Clean up utils.hub using the latest from hf_hub (#18857) 
* Clean up utils.hub using the latest from hf_hub

* Adapt test

* Address review comment

* Fix test
 2022-09-02 10:30:06 -04:00
NielsRogge
17981faf67 Add OWL-ViT to the appropriate section (#18867) 
Co-authored-by: Niels Rogge <nielsrogge@Nielss-MacBook-Pro.local>
 2022-09-02 15:59:25 +02:00
NielsRogge
c60dd98e87 [LayoutLM] Add clarification to docs (#18716) 
* Add clarification

* Add another clarification

* Apply suggestion

Co-authored-by: Niels Rogge <nielsrogge@Nielss-MacBook-Pro.local>
 2022-09-02 14:48:19 +02:00
OlivierDehaene
129d73294e Fix naming issue with ImageToText pipeline (#18864) 
Co-authored-by: Olivier Dehaene <olivier@huggingface.co>
 2022-09-02 07:55:30 -04:00
kmckiern
9b3eb81014 if learning rate is a tensor, get item (float) (#18861) 2022-09-02 07:46:31 -04:00
Steven Liu
142e12afb4 Split docs on modality (#18205) 
* update

* 🖍 add missing files

* 📝 add nested sections

* 🖍 align titles with tasks

* oops

* remove quotes from titles
 2022-09-01 15:19:11 -05:00
Ankur Goyal
23fab60b67 Pin revision for LayoutLMForQuestionAnswering and TFLayoutLMForQuestionAnswering tests (#18854) 
* Pin revision for tests

* Fixup

* Update revision in models

* Shorten revisions

Co-authored-by: Ankur Goyal <ankur@impira.com>
 2022-09-01 12:52:33 -04:00
OlivierDehaene
ddb69e5af8 Add Image To Text Generation pipeline (#18821) 
* Add Image2TextGenerationPipeline to supported pipelines

* Add Flax and Tensorflow support

* Add Flax and Tensorflow small tests

* Add default model for Tensorflow

* Add docstring

* Fix doc style

* Add tiny models for pytorch and flax

* Remove flax from pipeline.
Fix tests

* Use ydshieh/vit-gpt2-coco-en as a default for both PyTorch and Tensorflow

* Fix Tensorflow support

Co-authored-by: Olivier Dehaene <olivier@huggingface.co>
 2022-09-01 12:07:14 -04:00
Sylvain Gugger
c61f116b63 Tie weights after preparing the model in run_clm (#18855) 2022-09-01 12:06:56 -04:00
Cody Yu
1c381f3600 Cache results of is_torch_tpu_available() (#18777) 
* Cache results of is_torch_tpu_available()

* Update src/transformers/utils/import_utils.py

* Update src/transformers/utils/import_utils.py
 2022-09-01 11:45:33 -04:00
Sayak Paul
954e18ab97 TensorFlow MobileViT (#18555) 
* initial implementation.

* add: working model till image classification.

* add: initial implementation that passes intg tests.

Co-authored-by: Amy <aeroberts4444@gmail.com>

* chore: formatting.

* add: tests (still breaking because of config mismatch).

Coo-authored-by: Yih <2521628+ydshieh@users.noreply.github.com>

* add: corrected tests and remaning changes.

* fix code style and repo consistency.

* address PR comments.

* address Amy's comments.

* chore: remove from_pt argument.

* chore: add full-stop.

* fix: TFLite model conversion in the doc.

* Update src/transformers/models/mobilevit/modeling_tf_mobilevit.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/models/mobilevit/modeling_tf_mobilevit.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/models/mobilevit/modeling_tf_mobilevit.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/models/mobilevit/modeling_tf_mobilevit.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/models/mobilevit/modeling_tf_mobilevit.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* apply formatting.

* chore: remove comments from the example block.

* remove identation in the example.

Co-authored-by: Amy <aeroberts4444@gmail.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
 2022-09-01 10:35:15 -04:00