GitLab

* Do not prepare lr scheduler as it as the right number of steps

* Trigger CI

* Trigger CI

* Trigger CI

* Add fake comment

* Remove fake comment

* Trigger CI please!

* fix executable batch size issue

* fix

* undo

* Let's see if we can use the smallest possible dummies

* Make GPT-2's dummies a little longer

* Just use (1,2) as the default shape

* Update other dummies in sync

* Correct imports for Keras 2.13

* Shrink the Wav2Vec2 dummies

* fix skip modules test

* oops

* address comments

Fix is_optimum_neuron_available

add `safe_serialization` in push_to_hub

* support PEFT models when saving the model using trainer

* fixup

* Add support for non-rust implemented tokenization for `__getitem__` method.

* Update for error message on adding new sub-branch for `__item__` method.

---------
Co-authored-by: liuyang17 <liuyang17@zhihu.com>

* [Wav2Vec2] Fix torch srcipt

* fix more

* Fix model load when it has both code on the Hub and locally

* Add input check with timeout

* Add tests

* Apply suggestions from code review
Co-authored-by: Lysandre Debut <lysandre.debut@reseau.eseo.fr>

* Some non-saved stuff

* Add feature extractors

* Add image processor

* Add model

* Add processor and tokenizer

* Reduce timeout

---------
Co-authored-by: Lysandre Debut <lysandre.debut@reseau.eseo.fr>

* Fix device placement for model-parallelism in generate for encoder/decoders

* Remove debug statements

* Use new parametrization based weight norm if available

See https://github.com/pytorch/pytorch/pull/103001

Signed-off-by: Edward Z. Yang <ezyang@meta.com>

* handle copies
Signed-off-by: Edward Z. Yang <ezyang@meta.com>

* black
Signed-off-by: Edward Z. Yang <ezyang@meta.com>

---------
Signed-off-by: Edward Z. Yang <ezyang@meta.com>

* A fun new PR where I break the entire codebase again

* A fun new PR where I break the entire codebase again

* Handle cross-attention

* Move calls to model(model.dummy_inputs) to the new build() method

* Seeing what fails with the build context thing

* make fix-copies

* Let's see what fails with new build methods

* Fix the pytorch crossload build calls

* Fix the overridden build methods in vision_text_dual_encoder

* Make sure all our build methods set self.built or call super().build(), which also sets it

* make fix-copies

* Remove finished TODO

* Tentatively remove unneeded (?) line

* Transpose b in deberta correctly and remove unused threading local

* Get rid of build_with_dummies and all it stands for

* Rollback some changes to TF-PT crossloading

* Correctly call super().build()

* Make the default dummies (2, 2) instead of (3, 3)

* Fix for Funnel

* Actually fix Funnel

* Add test_backbone for convnext

* Add TimmBackbone model

* Add check for backbone type

* Tidying up - config checks

* Update convnextv2

* Tidy up

* Fix indices & clearer comment

* Exceptions for config checks

* Correclty update config for tests

* Safer imports

* Safer safer imports

* Fix where decorators go

* Update import logic and backbone tests

* More import fixes

* Fixup

* Only import all_models if torch available

* Fix kwarg updates in from_pretrained & main rebase

* Tidy up

* Add tests for AutoBackbone

* Tidy up

* Fix import error

* Fix up

* Install nattan in doc_test_job

* Revert back to setting self._out_xxx directly

* Bug fix - out_indices mapping from out_features

* Fix tests

* Dont accept output_loading_info for Timm models

* Set out_xxx and don't remap

* Use smaller checkpoint for test

* Don't remap timm indices - check out_indices based on stage names

* Skip test as it's n/a

* Apply suggestions from code review
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Cleaner imports / spelling is hard

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

Prevent ZeroDivisionError if evaluation is too quick

Co-authored-by: Andrey Voynov <avoin@google.com>

* Add check for tied parameters

* Fix style

* fix style

* Fix versioning

* Change if to elif

tensorboard callback no longer adds hparams

* fix wrong broadcast axis of attention mask in visual encoder

* fix slow tests

---------
Co-authored-by: younesbelkada <younesbelkada@gmail.com>

expose safe_serialization argument of PreTrainedModel and TFPreTrainedModel in the save_pretrained of the pipeline api
Co-authored-by: Yessen Kanapin <yessen@deepinfra.com>

add check loop over extra content

* fix trainer slow tests

* commit 2

* Add an option to reduce compile() console spam

* Add annotations to the example scripts

* Add notes to the quicktour docs as well

* minor fix

Trainer: fixed KeyError on evaluate for ReduceLROnPlateau
Co-authored-by: Claudius Kienle <claudius.kienle@artiminds.com>

* generated code from add-new-model-like

* Add code for modeling, config, and weight conversion

* add tests for image-classification, update modeling and config

* add code, tests for semantic-segmentation

* make style, make quality, make fix-copies

* make fix-copies

* Update modeling_mobilevitv2.py

fix bugs

* Update _toctree.yml

* update modeling, config

fix bugs

* Edit docs - fix bug MobileViTv2v2 -> MobileViTv2

* Update mobilevitv2.mdx

* update docstrings

* Update configuration_mobilevitv2.py

make style

* Update convert_mlcvnets_to_pytorch.py

remove unused options

* Update convert_mlcvnets_to_pytorch.py

make style

* Add suggestions from code review
Co-Authored-By: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* make style, make quality

* Add suggestions from code review
Co-Authored-By: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Add suggestions from code review

Remove MobileViTv2ImageProcessor
Co-Authored-By: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* make style

* Add suggestions from code review

Rename MobileViTv2 -> MobileViTV2
Co-Authored-By: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Add suggestions from code review
Co-Authored-By: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Update modeling_mobilevitv2.py

make style

* Update serialization.mdx

* Update modeling_mobilevitv2.py

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

* add fine-tuned with adapter layer

* Add set_target_lang to tokenizer

* Implement load adapter

* add tests

* make style

* Apply suggestions from code review

* Update src/transformers/models/wav2vec2/tokenization_wav2vec2.py

* make fix-copies

* Apply suggestions from code review

* make fix-copies

* make style again

* mkae style again

* fix doc string

* Update tests/models/wav2vec2/test_tokenization_wav2vec2.py

* Apply suggestions from code review

* fix

* Correct wav2vec2 adapter

* mkae style

* Update src/transformers/models/wav2vec2/modeling_wav2vec2.py
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>

* add more nice docs

* finish

* finish

* Apply suggestions from code review
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Apply suggestions from code review
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Apply suggestions from code review

* all finish

---------
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Fix 'ReduceLROnPlateau' object has no attribute 'get_last_lr'

* fix style

use _make_causal_mask in clip models

* Modify device map behavior for 4/8 bits model

* Remove device_map arg for training 4/8 bit model

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

* Add Exceptions

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

* Fix formatting

* Get current device with accelerate

* Revert "Get current device with accelerate"

This reverts commit 46f00799103bbe15bd58762ba029aab35363c4f7.

* Fix Exception

* Modify quantization doc

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

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

Make TF inv_freq non-trainable like PyTorch

rename encode input to match docstring

consistentcy