GitLab

* Add `OPTForQuestionAnswering`

- added `OPTForQuestionAnswering` class based on `BloomForQuestionAnswering`
- added `OPTForQuestionAnswering` in common tests
- all common tests pass
- make fixup done

* added docstrings for OPTForQuestionAnswering

* Fix docstrings for OPTForQuestionAnswering

The sequence_masked variable is actually the part of the sequence that is kept unmasked for the encoder. This commit renames the variable.

* Remove dependency of Roberta in Blenderbot

* Move Copied from statements to each method of the Roberta classes

* Remove copied from line for mask_token.setter

* update output from example in docs

* Add onnx support for VisionEncoderDecoder

* Add onnx support for VisionEncoderDecoder

* Removed unused import

* Rename encoder hidden state
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

* Update docstrings and removed redundant code

* Added test function for enc-dec models

* Update doc string text
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

* fixed code style
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

* Leverage hfh for move cache

* Style

Fix link typo of the following content.
PyTorch version, Trainer
PyTorch version, no Trainer

* fix marianMT convertion to onnx

* Update src/transformers/onnx/convert.py
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

* Update src/transformers/onnx/convert.py
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

* Fixed duplicated line (paragraph #83) @omarespejel @sgugger

* Datasets map denomination fixed (paragraph 42)

* remove RobertaConfig inheritance from MarkupLMConfig

* Update src/transformers/models/markuplm/configuration_markuplm.py

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

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

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

* Rework pipeline tests

* Try to fix Flax tests

* Try to put it before

* Use a new decorator instead

* Remove ignore marker since it doesn't work

* Filter pipeline tests

* Woopsie

* Use the fitlered list

* Clean up and fake modif

* Remove init

* Revert fake modif

- Fixes the image segmentation pipeline test failures caused by changes to the postprocessing methods of supported models
- Updates the ImageSegmentationPipeline tests
- Improves docs, adds 'task' argument to optionally perform semantic, instance or panoptic segmentation

* Copied all the code required from transformers.models.bert.modeling_bert to here

* Fixed styling issues

* Reformatted copied names with Model specific name.

* Reverted BertEncoder part as there is already a class called BertGenerationEncoder

* Added prefixes in missing places.
Co-authored-by: vishwaspai <vishwas.pai@emplay.net>

* camembert tf version independent

* fixup

* fixup, all working

* remove comments

* Adding copied from roberta
Co-authored-by: Mustapha AJEGHRIR <mustapha.ajeghrir@kleegroup.com>

* Copied from BertTokenizer() in tokenization_bert

* Added BasicTokenizer and WordPieceTokenizer Class

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

* Added copied from comments for basicTokenizer and WordPieceTokenizer

* Updated the comments for the tokenizerClasses

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

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

* Formatted tokenization_electra with `make style`

* Fix repo inconsistencies

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

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

* removed dependency from bart(slow)

* removed dependency from bart(slow)

* adding copying comments (copied from bart to led)

* updated led docstring

* updated led docstring

* removed dependency from Bart (fast)

* replaced bart with LED in docstrings

* complying flake8

* added more copy comments

* fixing copying comments

* added comments back

* fix copy comments

* fixing copied from comments

* fixing copied from comments

* add first generation tutorial

* uP

* [Clip] Add text model to device map

Co-authored-by: harry7337 <hari.8jan@gmail.com>

* Remove dependency of Bert from Squeezebert tokenizer

* run style corrections

* update copies from BertTokenizers

* Update changes and style to Squeezebert files

* update copies for bert-fast

* update feature extractor params

* update attention mask handling

* validate onnx models with a different input geometry than saved with

* only test working features for now

* simpler test skipping

* rm TODO

* expose batch_size/seq_length on vit

* skip certain name, feature, framework parameterizations known to fail validation

* Trigger CI

* Trigger CI

* edit: casting attention_mask to long in DataCollatorCTCWithPadding

* edit: casting attention_mask to long in DataCollatorCTCWithPadding

* Add ZeroShotObjectDetectionPipeline (#18445)

* Add AutoModelForZeroShotObjectDetection task

This commit also adds the following

- Add explicit _processor method for ZeroShotObjectDetectionPipeline.
  This is necessary as pipelines don't auto infer processors yet and
  `OwlVitProcessor` wraps tokenizer and feature_extractor together, to
  process multiple images at once

- Add auto tests and other tests for ZeroShotObjectDetectionPipeline

* Add AutoModelForZeroShotObjectDetection task

This commit also adds the following

- Add explicit _processor method for ZeroShotObjectDetectionPipeline.
  This is necessary as pipelines don't auto infer processors yet and
  `OwlVitProcessor` wraps tokenizer and feature_extractor together, to
  process multiple images at once

- Add auto tests and other tests for ZeroShotObjectDetectionPipeline

* Add batching for ZeroShotObjectDetectionPipeline

* Fix doc-string ZeroShotObjectDetectionPipeline

* Fix output format: ZeroShotObjectDetectionPipeline

* HF <-> megatron checkpoint conversion handling reshaping from different tensor and parallel sizes

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

* addressing comments

* add doc strings and  🐛

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

* Added type hints for TF: TransfoXL
* Added type hints for TF: TransfoXL

* Change type hints for training

* Change type hints for training

* swin transformer onnx support

* Updated image dimensions as dynamic
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>
Co-authored-by: lewtun <lewis.c.tunstall@gmail.com>

* Update modeling_tf_xlm.py

* Updates

* Update src/transformers/models/xlm/modeling_tf_xlm.py

* Update src/transformers/models/xlm/modeling_tf_xlm.py

* Update src/transformers/models/xlm/modeling_tf_xlm.py

* Update src/transformers/models/xlm/modeling_tf_xlm.py

* Update src/transformers/models/xlm/modeling_tf_xlm.py
Co-authored-by: Matt <Rocketknight1@users.noreply.github.com>

* ConvBert

* added comment

* Updated

* Final_updates

* Update tokenization_convbert.py

* Update tokenization_convbert_fast.py

* Update tokenization_convbert.py

* Update tokenization_convbert.py

* Update tokenization_convbert_fast.py

* Update tokenization_convbert.py

* Update tokenization_convbert_fast.py

* Updates

* Updates

* Updated

* Final Updates

The link to https://github.com/vasudevgupta7/bigbird is vulnerable to repojacking (it redirects to the orignial project that changed name), you should change the link to the current name of the project. if you won't change the link, an attacker can open the linked repository and attacks users that trust your links

This PR aims to rectify the discrepancy between the training performances of HF and Timm ViT implementations.

- Initializes torch and flax ViT dense layer weights with trunc_normal instead of normal (consistent with the TF implementation.
- Initializes cls_token and positional_embeddings with trunc_normal
- Updates DeiT copy to reflect the changes

* Fix pipeline tests for Roberta-like tokenizers

* Fix fix