GitLab

* fix

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

* fix

* fix

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

* make SpeechT5 model by copying Wav2Vec2

* add paper to docs

* whoops added docs in wrong file

* remove SpeechT5Tokenizer + put CTC back in the name

* remove deprecated class

* remove unused docstring

* delete SpeechT5FeatureExtractor, use Wav2Vec2FeatureExtractor instead

* remove classes we don't need right now

* initial stab at speech encoder prenet

* add more speech encoder prenet stuff

* improve SpeechEncoderPrenet

* add encoder (not finished yet)

* add relative position bias to self-attention

* add encoder CTC layers

* fix formatting

* add decoder from BART, doesn't work yet

* make it work with generate loop

* wrap the encoder into a speech encoder class

* wrap the decoder in a text decoder class

* changed my mind

* changed my mind again ;-)

* load decoder weights, make it work

* add weights for text decoder postnet

* add SpeechT5ForCTC model that uses only the encoder

* clean up EncoderLayer and DecoderLayer

* implement _init_weights in SpeechT5PreTrainedModel

* cleanup config + Encoder and Decoder

* add head + cross attention masks

* improve doc comments

* fixup

* more cleanup

* more fixup

* TextDecoderPrenet works now, thanks Kendall

* add CTC loss

* add placeholders for other pre/postnets

* add type annotation

* fix freeze_feature_encoder

* set padding tokens to 0 in decoder attention mask

* encoder attention mask downsampling

* remove features_pen calculation

* disable the padding tokens thing again

* fixup

* more fixup

* code review fixes

* rename encoder/decoder wrapper classes

* allow checkpoints to be loaded into SpeechT5Model

* put encoder into wrapper for CTC model

* clean up conversion script

* add encoder for TTS model

* add speech decoder prenet

* add speech decoder post-net

* attempt to reconstruct the generation loop

* add speech generation loop

* clean up generate_speech

* small tweaks

* fix forward pass

* enable always dropout on speech decoder prenet

* sort declaration

* rename models

* fixup

* fix copies

* more fixup

* make consistency checker happy

* add Seq2SeqSpectrogramOutput class

* doc comments

* quick note about loss and labels

* add HiFi-GAN implementation (from Speech2Speech PR)

* rename file

* add vocoder to TTS model

* improve vocoder

* working on tokenizer

* more better tokenizer

* add CTC tokenizer

* fix decode and batch_code in CTC tokenizer

* fix processor

* two processors and feature extractors

* use SpeechT5WaveformFeatureExtractor instead of Wav2Vec2

* cleanup

* more cleanup

* even more fixup

* notebooks

* fix log-mel spectrograms

* support reduction factor

* fixup

* shift spectrograms to right to create decoder inputs

* return correct labels

* add labels for stop token prediction

* fix doc comments

* fixup

* remove SpeechT5ForPreTraining

* more fixup

* update copyright headers

* add usage examples

* add SpeechT5ProcessorForCTC

* fixup

* push unofficial checkpoints to hub

* initial version of tokenizer unit tests

* add slow test

* fix failing tests

* tests for CTC tokenizer

* finish CTC tokenizer tests

* processor tests

* initial test for feature extractors

* tests for spectrogram feature extractor

* fixup

* more fixup

* add decorators

* require speech for tests

* modeling tests

* more tests for ASR model

* fix imports

* add fake tests for the other models

* fixup

* remove jupyter notebooks

* add missing SpeechT5Model tests

* add missing tests for SpeechT5ForCTC

* add missing tests for SpeechT5ForTextToSpeech

* sort tests by name

* fix Hi-Fi GAN tests

* fixup

* add speech-to-speech model

* refactor duplicate speech generation code

* add processor for SpeechToSpeech model

* add usage example

* add tests for speech-to-speech model

* fixup

* enable gradient checkpointing for SpeechT5FeatureEncoder

* code review

* push_to_hub now takes repo_id

* improve doc comments for HiFi-GAN config

* add missing test

* add integration tests

* make number of layers in speech decoder prenet configurable

* rename variable

* rename variables

* add auto classes for TTS and S2S

* REMOVE CTC!!!

* S2S processor does not support save/load_pretrained

* fixup

* these models are now in an auto mapping

* fix doc links

* rename HiFiGAN to HifiGan, remove separate config file

* REMOVE auto classes

* there can be only one

* fixup

* replace assert

* reformat

* feature extractor can process input and target at same time

* update checkpoint names

* fix commit hash

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

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

* force `memory_efficient_backward=True`

* enhancements

- trainer support
- add new flag

* some changes

- internal changes in `Trainer`
- small refactor

* make quality

* Fixes

- add new testing util
- add new test
- change test in Trainer

* fix CI test

* educate users on how to ft 8bit models

* more checks

* fix `logger` error

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

* adapt from review

* fix

* add comment

* use return instead

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

* [FIX] path for Graphormer checkpoint

* [FIX] Test suite for graphormer

* [FIX] Update graphormer default num_classes

* gelu_python_tanh

* rename

* Version check, add test

* Pr comment

* Fix the input embeds issue with tests

* Fix black and isort issue

* Clean up tests

* Add slow tag to the test introduced

* Incorporate PR feedbacks

* Bump onnx in /examples/research_projects/decision_transformer

Bumps [onnx](https://github.com/onnx/onnx) from 1.11.0 to 1.13.0.
- [Release notes](https://github.com/onnx/onnx/releases)
- [Changelog](https://github.com/onnx/onnx/blob/main/docs/Changelog.md)
- [Commits](https://github.com/onnx/onnx/compare/v1.11.0...v1.13.0

)

---
updated-dependencies:
- dependency-name: onnx
  dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>

* adapt

* finish

* Update examples/research_projects/decision_transformer/requirements.txt

* up

* add tests

* Apply suggestions from code review
Co-authored-by: Lucain <lucainp@gmail.com>
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

* fix test

---------
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Lucain <lucainp@gmail.com>
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>

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

Generate: fix TF XLA tests on models with `max_position_embeddings` or `max_target_positions` (#21389)

* First draft

* Add initial draft of conversion script

* Convert all weights

* Fix config

* Add image processor

* Fix DetaImageProcessor

* Run make fix copies

* Remove timm dependency

* Fix dummy objects

* Improve loss function

* Remove conv_encoder attribute

* Update conversion scripts

* Improve postprocessing + docs

* Fix copied from statements

* Add tests

* Improve postprocessing

* Improve postprocessing

* Update READMEs

* More improvements

* Fix rebase

* Add is_torchvision_available

* Add torchvision dependency

* Fix typo and README

* Fix bug

* Add copied from

* Fix style

* Apply suggestions

* Fix thanks to @ydshieh

* Fix another dependency check

* Simplify image processor

* Add scipy

* Improve code

* Add threshold argument

* Fix bug

* Set default threshold

* Improve integration test

* Add another integration test

* Update setup.py

* Address review

* Improve deformable attention function

* Improve copied from

* Use relative imports

* Address review

* Replace assertions

* Address review

* Update dummies

* Remove dummies

* Address comments, update READMEs

* Remove custom kernel code

* Add image processor tests

* Add requires_backends

* Add minor comment

* Update scripts

* Update organization name

* Fix defaults, add doc tests

* Add id2label for object 365

* Fix tests

* Update task guide

[FIX] path for Graphormer checkpoint

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

* rework pipeline tests

* run pipeline tests

* fix

* fix

* fix

* revert the changes in get_test_pipeline() parameter list

* fix expected error message

* skip a test

* clean up

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

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

* another patch

* fix timestamp test modeling

* let it be negative when the token is None

remove max_length=None
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* update

* update all

* clean up

* make quality

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

* Accept a batched tensor of images as input

* Add to all image processors

* Update oneformer

* add small patch

* update tests, forced decoder ids is not prioritary against generation config

* fix two new tests

* Commit with BTModel and latest HF code

* Placeholder classes for BTForMLM and BTForITR

* Importing Bert classes from transformers

* Removed objectives.py and dist_utils.py

* Removed swin_transformer.py

* Add image normalization, BridgeTowerForImageAndTextRetrieval

* Add center_crop

* Removing bert tokenizer and LCI references

* Tested config loading from HF transformers hub

* Removed state_dict updates and added path to hub

* Enable center crop

* Getting image_size from config, renaming num_heads and num_layers

* Handling max_length in BridgeTowerProcessor

* Add BridgeTowerForMaskedLM

* Add doc string for BridgeTowerConfig

* Add doc strings for BT config, processor, image processor

* Adding docs, removed swin

* Removed convert_bridgetower_original_to_pytorch.py

* Added doc files for bridgetower, removed is_vision

* Add support attention_mask=None and BridgeTowerModelOutput

* Fix formatting

* Fixes with 'make style', 'make quality', 'make fixup'

* Remove downstream tasks from BridgeTowerModel

* Formatting fixes, add return_dict to BT models

* Clean up after doc_test

* Update BTModelOutput return type, fix todo in doc

* Remove loss_names from init

* implement tests and update tuples returned by models

* Add image reference to bridgetower.mdx

* after make fix-copies, make fixup, make style, make quality, make repo-consistency

* Rename class names with BridgeTower prefix

* Fix for image_size in BTImageProcessor

* implement feature extraction bridgetower tests

* Update image_mean and image_std to be list

* remove unused import

* Removed old comments

* Rework CLIP

* update config in tests followed config update

* Formatting fixes

* Add copied from for BridgeTowerPredictionHeadTransform

* Update bridgetower.mdx

* Update test_feature_extraction_bridgetower.py

* Update bridgetower.mdx

* BridgeTowerForMaskedLM is conditioned on image too

* Add BridgeTowerForMaskedLM

* Fixes

* Call post_init to init weights

* Move freeze layers into method

* Remove BTFeatureExtractor, add BT under multimodal models

* Remove BTFeatureExtractor, add BT under multimodal models

* Code review feedback - cleanup

* Rename variables

* Formatting and style to PR review feedback

* Move center crop after resize

* Use named parameters

* Style fix for modeling_bridgetower.py

* Update docs/source/en/model_doc/bridgetower.mdx
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Update docs/source/en/model_doc/bridgetower.mdx
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Update docs/source/en/model_doc/bridgetower.mdx
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

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

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

* Update docs/source/en/model_doc/bridgetower.mdx
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>

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

* Rename config params, copy BERT classes, clean comments

* Cleanup irtr

* Replace Roberta imports, add BTTextConfig and Model

* Update docs, add visionconfig, consistent arg names

* make fixup

* Comments for forward in BTModel and make fixup

* correct tests

* Remove inconsistent roberta copied from

* Add BridgeTowerTextModel to dummy_pt_objects.py

* Add BridgeTowerTextModel to IGNORE_NON_TESTED

* Update docs for BT Text and Vision Configs

* Treat BridgeTowerTextModel as a private model

* BridgeTowerTextModel as private

* Run make fix-copies

* Adding BTTextModel to PRIVATE_MODELS

* Fix for issue with BT Text and Image configs

* make style changes

* Update README_ja.md

Add から to BridgeTower's description

* Clean up config, .mdx and arg names

* Fix init_weights. Remove nn.Sequential

* Formatting and style fixes

* Re-add tie_word_embeddings in config

* update test implementation

* update style

* remove commented out

* fix style

* Update README with abs for BridgeTower

* fix style

* fix mdx file

* Update bridgetower.mdx

* Update img src in bridgetower.mdx

* Update README.md

* Update README.md

* resolve style failed

* Update _toctree.yml

* Update README_ja.md

* Removed mlp_ratio, rename feats, rename BTCLIPModel

* Replace BTCLIP with BTVisionModel,pass in vision_config to BTVisionModel

* Add test_initialization support

* Add support for output_hidden_states

* Update support for output_hidden_states

* Add support for output_attentions

* Add docstring for output_hidden_states

* update tests

* add bridgetowervisionmodel as private model

* rerun the PR test

* Remove model_type, pass configs to classes, renames

* Change self.device to use weight device

* Remove image_size

* Style check fixes

* Add hidden_size and num_hidden_layers to BridgeTowerTransformer

* Update device setting

* cosmetic update

* trigger test again

* trigger tests again

* Update test_modeling_bridgetower.py

trigger tests again

* Update test_modeling_bridgetower.py

* minor update

* re-trigger tests

* Update docs/source/en/model_doc/bridgetower.mdx
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Remove pad, update max_text_len, doc cleanup, pass eps to LayerNorm

* Added copied to, some more review feedback

* make fixup

* Use BridgeTowerVisionEmbeddings

* Code cleanup

* Fixes for BridgeTowerVisionEmbeddings

* style checks

* re-tests

* fix embedding

* address comment on init file

* retrigger tests

* update import prepare_image_inputs

* update test_image_processing_bridgetower.py to reflect test_image_processing_common.py

* retrigger tests
Co-authored-by: Shaoyen Tseng <shao-yen.tseng@intel.com>
Co-authored-by: Tiep Le <tiep.le@intel.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>
Co-authored-by: Tiep Le <97980157+tileintel@users.noreply.github.com>

* update values

* update values

* update values

* Update tests/models/oneformer/test_modeling_oneformer.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

Moving to cleaner tokenizer version.

* update whisper logit processor

* add generate for whisper

* remove part of the whisper specific code from pipeline

* update logit processes

* major update

* enforce first timestamp

* update generate

* add more tests

* update new decoding strategy

* Apply suggestions from code review

* update docstring

* fixup

* default config will not have multilingual ar

* update expected tokenizer size, see pull on the hub for whisper-tiny

* Fixing the pipeline with image processor.

* Update the slow test.

* Using only the first image processor.

* Include exclusion mecanism for Image processor.

* Do not handle Gitconfig, deemed as a bug.

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

* Remove `conversational` changes. They are not supposed to be here.

* Address first row of comments.

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

* Add test

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

* add additional kwargs handling

* fix issue when serializing

* correct order of kwargs removal for serialization in from dict

* add `dict_torch_dtype_to_str` in case a dtype is needed for generation

* add condition when adding the kwargs : not from config

* Add comment based on review
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* add test function

* default None when poping arg
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* attempts to fix:

- upcast input for `T5DenseActDense`
- add the condition `self.wo.weight.dtype != torch.int8`
- added tests on `test/mixed_int8`
- `make fixup`

* fix ci test

* fix instance segmentation post processing

* add Mask2FormerImageProcessor

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

* Replace reduce_labels with do_reduce_labels

* Replace only for __init__ and preprocess

* Add class properties with warnings

* Update tests

* use streaming dataset

* fix whisper's test

* add rescale argument to chunk_iter