GitLab

* fix confusing PIL instructions

As stated in the documentation
[here](https://pillow.readthedocs.io/en/stable/handbook/image-file-formats.html?highlight=pdf#write-only-formats

),
PIL can only write PDF's, not read them. Remove references to reading
PDF's via PIL from this page to avoid confusion.

* mention PDF in doc examples using PIL
Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>

* Be explicit: PDFs must be converted to images

* fix formatting
Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>

* ✨ create concept guide section

* 🖍 make fixup

* 🖍

 apply feedback
Co-authored-by: Steven <stevhliu@gmail.com>

* Initial commit

* Add TFGPTJModel

* Fix a forward pass

* Add TFGPTJCausalLM

* Add TFGPTJForSequenceClassification

* Add TFGPTJForQuestionAnswering

* Fix docs

* Deal with TF dynamic shapes

* Add Loss parents to models

* Adjust split and merge heads to handle 4 and 5-dim tensors

* Update outputs for @tooslow tests

* Add ONNX support for Blenderbot

* Add BlenderbotSmall ONNX configuration

* Update serialization table

* Sharded checkpoint support

* Handle distant sharded checkpoints

* Add tests

* TODO is done

* Apply suggestions from code review
Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>

* Fix docstring

* Add example and format

* Address review comments

* More review comments

* End of merge

* Revert unintentional change

* VsCode what did you do?

* Style

* Changes

* Address final comments

* Quality

* Moar tests

* Move import beneath is_pt_available
Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>

* Big file_utils cleanup

* This one still needs to be treated separately

Co-authored-by: Niels Rogge <nielsrogge@Nielss-MacBook-Pro.local>

* Added type hints for PyTorch T5 model

* removed a type hint

* ran make style

* added type hints for ibert pytorch

* added type hints for lxmert pytorch

* removed kwargs type hint and fixed arguments order

* variable naming for Distilbert model

* adding unpack inputs at top

* make style/quality
Co-authored-by: matt <rocketknight1@gmail.com>

* Fix doc links in README

* Fix name

* Fix links in READMEs and doc index

* Error if there is something wrong so the CI knows

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* update PT/Flax equivalence tests on PT side

* overwrite check_outputs in BigBirdModelTest
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* Add missing type hints for ConvBERT flavored models.

* Update src/transformers/models/convbert/modeling_convbert.py
Co-authored-by: Matt <Rocketknight1@users.noreply.github.com>

* Type hints and TF decorator added

* Re-add XLA generation method

* Re-add lines that were deleted by conflicting updates

* Re-add lines that were deleted by conflicting updates

* Re-add lines that were deleted by conflicting updates
Co-authored-by: matt <rocketknight1@gmail.com>

* fix

* update the expected value in test_fast_integration
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* Update readme with how to train offline and fix BPE command

* Update examples/research_projects/codeparrot/README.md
Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* Update examples/research_projects/codeparrot/README.md
Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* Update examples/research_projects/codeparrot/README.md
Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* Update examples/research_projects/codeparrot/README.md
Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>
Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* update doc examples for TFRoberta

* fix style

* fix style

* use TF ckpt

* apply suggestion

* add the code file to test here

* fix style
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* [Doctests] Make roberta-like meaningfull

* correct

* final correct

* Trigger test

* make style

* apply suggestion from sylvain

* Make BigBird model compatiable to fp16 dtype.

* Use tree_map instead of map

* Reformat the code

* Fix import order

* Convert masks to the correct dtype

* Fix format issue

* Address comments.

* Update docs/README.md
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* add GPT-J ONNX config to Transformers

* remove token-classification features mapping
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

* add question-answering features mapping
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

* add GPT2 config init to GPT2 config + copie shebang for fix-copies
Co-authored-by: ChainYo <t.chaigneau.tc@gmail.com>
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

* Created the Decision Transformer Modle

* updating tests, copy to other machine

* Added last hidden size to Decision Transformer modelling outputs

* Removed copy of original DT file

* made a temporary change to gpt2 to have it conform with the Decision Transformer version

* Updated tests

* Ignoring a file used to test the DT model

* added comments to config file

* added comments and argument descriptions to decision transformer file

* Updated doc

* Ran "make style"

* Remove old model imports

* Removed unused imports, cleaned up init file

* Update docs/source/model_doc/decision_transformer.mdx

added my username
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

* Reverted changes made to gpt2

* Removed datasets submodule

* Update the modeling outputs to include gpt2 attentions, hidden states and last hidden states

* Added support for return of hidden states, attentions and return dict of gpt2 model.

* Updat...

* Make Transformers use cache files when hf.co is down

* Fix tests

* Was there a random circleCI failure?

* Isolate patches

* Style

* Comment out the failure since it doesn't fail anymore

* Better comment

* Swap inequalities

* Update src/transformers/trainer_callback.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/trainer_callback.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* revert tf gpt2

* add test for unpack_inputs and fix test case

* add changes to vision encoder decoder

* added type hints for mbart tensorflow tf implementation

* Adding missing type hints for mBART model 

Tensorflow Implementation model added with missing type hints

* Missing Type hints - correction

For TF model

* Code fixup using make quality tests

* Hint types - typo error

* make fix-copies and make fixup

* type hints

* updated files
Co-authored-by: matt <rocketknight1@gmail.com>

* Initial commit

* Reversed signs, adjusted log entery.

* Check only when

* Cleanup checks

* Only trigger if we want to eval

* Run

* Move changes to callback

* [FlaxBart] make sure no grads are computed an bias

* correct all other seq2seq models