GitLab

* fix BLOOM ONNX config
- `value` params have `seq_len` as their 2nd axe as opposed to other models which have it as 3rd
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

* Add doc tests

* Make it more consistent
Co-authored-by: Niels Rogge <nielsrogge@Nielss-MacBook-Pro.local>

* [Whisper] Don't return attention mask in feat extractor

* remove attention mask from test

* fix failing tests

* quality

* Cast masks to np.unit8 before converting to PIL.Image.Image

* Update tests

* Fixup

* [Doctest] Add `configuration_bigbird_pegasus.py` and `configuration_big_bird`

[Doctest] Re-style `configuration_big_bird.py`

* [Doctest] One python instruction per line

* [Doctest] Fix styling

* [Doctest] More styling fixes

* Added type hints for TF: XLNet

* Added type hints for TF: XLNet

* Added type hints for TF: XLNet

* Added type hints for TF: XLNet

* Added type hints for TF: XLNet

* Added type hints for TF: XLNet

* Add type hints for XLnet (TF)
* Added type hints for XLnet (TF)

* Update src/transformers/models/xlnet/modeling_tf_xlnet.py

* yoso config for doctest

* Revert "yoso config for doctest"

This reverts commit eae128d6f1b3631b676ffbcc181390e338819bd1.

* add configurations_blenderbot_small.py for doctests

* yoso config for doctest

* Revert "yoso config for doctest"

This reverts commit eae128d6f1b3631b676ffbcc181390e338819bd1.

* add configurations.blenderbot.py for doctests

* add configuration.blenderbot for doctest

* Add type hints for Lxmert (TF)

* Update src/transformers/models/lxmert/modeling_tf_lxmert.py
Co-authored-by: Emmanuel Lusenji <elusenji@Emmanuels-MacBook-Pro.local>
Co-authored-by: Matt <Rocketknight1@users.noreply.github.com>

* Update modeling_deberta_v2.py

* Update modeling_deberta_v2.py

* fix loading fairseq wav2vec2 pretrained weights

Specified fairseq task as "audio_pretraining" when loading fairseq weights,
since loading wav2vec2-xls-r weights fails if the task is unspecified.

Resolves: #19319

* fix style

* Compute true loss

* fixup

* final

* final

* final

* Update examples/flax/language-modeling/run_bart_dlm_flax.py
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>

* jax.tree_map => jax.tree_util.tree_map

* Compute true loss

* final

* fixup

* final

* final

* Update examples/flax/language-modeling/run_bart_dlm_flax.py
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>

* jax.tree_map => jax.tree_util.tree_map
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>

* ViT Config for doctest

* Added (with random weights) in the comment before model initialization line

* Added configuration_bert_generation.py to utils/documentation_tests.txt
Co-authored-by: vishwaspai <vishwas.pai@emplay.net>

* Fixing mobile bert configuration doctest

* Fixed build failures by removing empty line

* [Whisper] Freeze params of encoder

* add tests

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

* Build Push CI images also in a daily basis

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

* create the arange tensor on device for enabling CUDA-Graph at higher-performace for SD

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

* return None to avoid recursive call

* Give error

* Give error

* Add test

* More tests

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

* Whisper Config for doctest

* restyle fix