GitLab

* fix memory leak
* fix typos
* use singular last hidden state variable names
* eliminate double call to self.owlvit to return last hidden states
* eliminate 2nd call to self.vision_model in OwlViTModel

* Add security warning about from_pretrained() method

* Add sentence about malware scanner
Co-authored-by: Julien Chaumond <julien@huggingface.co>

* Add ONNX support

* Make height and width dynamic axes
Co-authored-by: Niels Rogge <nielsrogge@Nielss-MacBook-Pro.local>

* fix args for bram search decoding in generation utils

* fix missing PAD token in gpt2

* add PAD EOS change to TF

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

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

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

Add an option to `HfArgumentParser.parse_{dict,json_file}` to raise an Exception when there extra keys (#18692)

* Update parser to track unneeded keys, off by default

* Fix formatting

* Fix docstrings and defaults in HfArgparser

* Fix formatting

* Minor typo in GPT2 doc

* improve gpt2 label doc

* update dim of label in GPT2ForTokenClassification

* add change to tf

* Pin ffspec

* Typo

* Add num_channels attribute

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

fix bug: register_for_auto_class should be defined on TFPreTrainedModel instead of TFSequenceSummary (#18607)

* Check optimizer for device on TPU

* Typo

* oob performance improvement for cpu DDP
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>

* add is_psutil_available check
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>

* Fix loss condition in DetrHungarianMatcher

* Fix costs condition in YolosHungarianMatcher

* Add LayoutLMForQuestionAnswering model

* Fix output

* Remove TF TODOs

* Add test cases

* Add docs

* TF implementation

* Fix PT/TF equivalence

* Fix loss

* make fixup

* Fix up documentation code examples

* Fix up documentation examples + test them

* Remove LayoutLMForQuestionAnswering from the auto mapping

* Docstrings

* Add better docstrings

* Undo whitespace changes

* Update tokenizers in comments

* Fixup code and remove `from_pt=True`

* Fix tests

* Revert some unexpected docstring changes

* Fix tests by overriding _prepare_for_class
Co-authored-by: Ankur Goyal <ankur@impira.com>

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

* Improving the documentation for "word", within the pipeline.

* Quality.

* Added Doctest for Deberta Pytorch

* Added path in documentation test file

* Added docstrings for DebertaV2

* Revert "Added docstrings for DebertaV2"

This reverts commit 307185e62a21b3bd0923444cc8a8af1747fd2600.

* Added DebertaV2 Docstrings

* groupvit to onnx

* dynamic shape for pixel values dim

* onnx conversion for owlvit

* .T to .t()

* dynamic shapes for pixel values

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

* Run tests if skip condition not met

* Update comment - remove outdated ref to TF 2.8

Co-authored-by: Esben Toke Christensen <esben.christensen@visma.com>
Co-authored-by: Lasse Reedtz <lasse.reedtz@visma.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Add dtype

* Fix Swinv2 as well
Co-authored-by: Niels Rogge <nielsrogge@Nielss-MacBook-Pro.local>

* add docstring for BartForCausalLM

* doc-style fic

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

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

Co-authored-by: Lysandre Debut <hi@lysand.re>
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

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

* Fix broken link

* Trigger CI
Co-authored-by: Stas Bekman <stas@stason.org>

* Fix incomplete FlaxBert outputs

* fix big_bird electra roberta

[Wav2vec2 + LM Test] Improve wav2vec2 with lm tests and make torch version dependent for now (#18749)

* add first generation tutorial

* remove generation

* make version dependent expected values

* Apply suggestions from code review

* Update tests/models/wav2vec2_with_lm/test_processor_wav2vec2_with_lm.py

* fix typo

* add first generation tutorial

* VisionEnocderDecoder gradient checkpointing

* remove generation

* add tests