GitLab

* Fix skip_special_tokens process for Wav2Vec2CTCTokenizer._decode

* Fix skip_special_tokens for Wav2Vec2CTCTokenizer._decode

* Exclude pad_token filtering since it is used as CTC-blank token

* Add small test for skip_special_tokens

* Update decoding test for added new token

* add FA2 to o.g Musicgen

* make style

* add FA2 support to Musicgen Melody

* add generation FA2 tests to o.g Musicgen

* make style and fix copies

* add Musicgen to FA2 docs + deprecate list

* add sdpa supports to Musicgen's

* make style and fix copies

* refactor attention implementation arguments

* add Copied from to sdpa tests

* add copied form in sdpa tests melody

* add copied for FA2 generation tests

* add FA2 inference copied from

* make style

* fix issue with logit processor in beam search in Flax

* adding FlaxNoRepeatNGramLogitsProcessor class + unit test

* style correction and code verification

* add FlaxNoRepeatNGramLogitsProcessor to the test_processor_list and test_processor_list_jitted tests

* fix an issue where ngrams are banned only if they appear ==1 time + update description of get_previous_ngrams

* replace non-jit compatible masking of ngrams that are not yet generated with jittable version

* Revert "fix issue with logit processor in beam search in Flax"

This reverts commit 09b70d7e4dc32d0cc4db61af09a835a9cd238b50.

* add FlaxNoRepeatNGramLogitsProcessor to _get_logits_processor

* change the method of casting to boolean of banned tokens indices

* fix code style

* remove some useless operations + significantly faster computation of update indices using jax.lax.fori_loop

* remove useless loop iterations

* set some variables that were calculated and used multiple times

* fix format

fix bug

* Fix sinusoidal_embeddings in FlaubertModel

* Fix for Informer

* Fix for XLM

* Move sinusoidal emb for XLM

* Move sinusoidal emb for Flaubert

* Small cleanup

* Add comments on tests code copied from

* Add with Distilbert->

* fix bug and add tests

* nit

* otherway to get the cur len instead of attention mask

* more places where this might have been broken

* nit

* oups

* inputs_embeds vs input_embeds

* test generated outptus

* style

* nit

* fix

* skip failing biogpt

* update

* feedback

[tests] fix the wrong output in `ImageToTextPipelineTests.test_conditional_generation_llava` (#29975)

bug fix

* fix copies

* nit

* style

* Update utils/check_copies.py

* fix FA2 tests

* refactor inference test name

* Start rework

* Fix failing test

* Include max

* Update src/transformers/trainer.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

---------
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

fix awq quant

* Update qwen2_moe.md

* update link of blogpost.

* fixup

---------
Co-authored-by: bozheng-hit <dsoul0621@gmail.com>

Fixes #29690

Co-authored-by: Alexander Jipa <azzhipa@amazon.com>

* improve: error message for best model metric

* update: raise warning instead of error

fix: rope_theta for open llama

* with with

* style

* fix

* revert for qwen2

* revert for qwen2

* update

* update

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

Trainer with PyTorch now requires accelerate to be installed.

Partly resolves huggingface/transformers#29174

* fix

* fix test

* style

* nit

* rather rely on concert token to id

* fix quality

* Update src/transformers/convert_slow_tokenizer.py

Fix doc issue in DebertaV2Config class
Co-authored-by: Vinayakk Garg <vigar@akamai.com>

* fi xbc?

* nit

* add gradient_accumulation_kwargs to AcceleratorConfig

* add suggestions from @muellerzr to docstrings, new behavior and tests

* Documentation suggestions from @muellerz
Co-authored-by: Zach Mueller <muellerzr@gmail.com>

* addressed @muellerzr comments regarding tests and test utils

* moved accelerate version to top of file.

* @muellerzr's variable fix
Co-authored-by: Zach Mueller <muellerzr@gmail.com>

* address @amyeroberts. fix tests and docstrings

* address @amyeroberts additional suggestions

---------
Co-authored-by: Yu Chin Fabian Lim <flim@sg.ibm.com>
Co-authored-by: Zach Mueller <muellerzr@gmail.com>

[ `TokenizationLlama`] fix the way we convert tokens to strings to keep leading spaces 🚨 breaking fix (#29453)

* nit

* update test and fix test

* fixup

* nit

* update

* oups

* Update src/transformers/models/mamba/modeling_mamba.py
Co-authored-by: Lysandre Debut <hi@lysand.re>

---------
Co-authored-by: Lysandre Debut <hi@lysand.re>

* add hard rope scaling test

* make fixup

* quick rope scaling tests

* add copy statements

* add functions to get number of params which require grad, get optimizer group for parameters and get learning rates of param groups to trainer.py

* add tests and raise ValueError when optimizer is None

* add second layer to test and freeze its weigths

* check if torch is available before running tests

* use decorator to check if torch is available
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* fix test indentation
Co-authored-by: Zach Mueller <muellerzr@gmail.com>

---------
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
Co-authored-by: Zach Mueller <muellerzr@gmail.com>

* Safe import of LRScheduler

* Update src/transformers/trainer_pt_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* Update src/transformers/trainer_pt_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* Fix up

---------
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* replace torch.testing.assert_allclose by torch.testing.assert_close

* missing atol rtol

fix typo

* First commit to add flash attention 2 for GPT-2

* more improvements

* Make GPT2 pass tests and fixed Decison Transformers copies

* Fixed missing arg

* fix copies

* Added expected speedup

* Update src/transformers/models/gpt2/modeling_gpt2.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* Update src/transformers/models/gpt2/modeling_gpt2.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* Update src/transformers/models/gpt2/modeling_gpt2.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* Added test

* Fixed attn attribute

* Update docs/source/en/model_doc/gpt2.md
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* Update docs/source/en/model_doc/gpt2.md
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* Update Decision transformer attentions

* More updates

* Passing tests

* Fix copies

* Fix copies part 2

* Decision transformer updates

* Update src/transformers/models/gpt2/modeling_gpt2.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Fix copies

* Decision transformer not supporting flash attn

* Addressed comments

* Addressed comments

* Addressed comments

---------
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* add doc warning

* fix build pr

don't gather on pkv when using the trainer

* add some help

* style

This commit adds gate jitter to MixtralSparseMoeBlock's input data
before passing it through the MoE layer, if turned on.