GitLab

* fsdp+qlora related changes

* fixes

* Update quantization_config.py

* support fsdp+qlora and dsz3+qlora

* Update quantization_config.py

* Update modeling_utils.py

* Update modeling_utils.py

* Update modeling_utils.py

* Update modeling_utils.py

* Update modeling_utils.py

* Update modeling_utils.py

* handle fsdp+qlora and dsz3+qlora correctly while model loading

* fix param count

* quality

* fsdp related changes

* fsdp changes only when using LoRA/QLoRA

* add accelerate version check

* refactor, update min accelerate version and add tests

1. Update minimum accelerate version to 0.26.0
2. Clean the trainer wrt accelerate version checks
3. FSDP refactor and test for fsdp config
4. use `itemsize` instead of `dtype2bytes` dict

* fix test

* Address comments
Co-Authored-By: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>

* fix the conditional flag

* fix conditional flag

* address comments
Co-Authored-By: Zach Mueller <7831895+muellerzr@users.noreply.github.com>

---------
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>
Co-authored-by: Zach Mueller <7831895+muellerzr@users.noreply.github.com>

* Adds pretrained IDs directly in the tests

* Fix tests

* Fix tests

* Review!

* initial implementation of flash attention for gptj

* modify flash attention and overwrite test_flash_attn_2_generate_padding_right

* update flash attention support list

* remove the copy line in the `CodeGenBlock`

* address copy mechanism

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

* Add GPTJ attention classes

* add expected outputs in the gptj test

* Ensure repo consistency with 'make fix-copies'

---------
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* add tests for batching support

* Update src/transformers/models/fastspeech2_conformer/modeling_fastspeech2_conformer.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Update src/transformers/models/fastspeech2_conformer/modeling_fastspeech2_conformer.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Update tests/test_modeling_common.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Update tests/test_modeling_common.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Update tests/test_modeling_common.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* fixes and comments

* use cosine distance for conv models

* skip mra model testing

* Update tests/models/vilt/test_modeling_vilt.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* finzalize  and make style

* check model type by input names

* Update tests/models/vilt/test_modeling_vilt.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* fixed batch size for all testers

* Revert "fixed batch size for all testers"

This reverts commit 525f3a0a058f069fbda00352cf202b728d40df99.

* add batch_size for all testers

* dict from model output

* do not skip layoutlm

* bring back some code from git revert

* Update tests/test_modeling_common.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Update tests/test_modeling_common.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* clean-up

* where did minus go in tolerance

* make whisper happy

* deal with consequences of losing minus

* deal with consequences of losing minus

* maskformer needs its own test for happiness

* fix more models

* tag flaky CV models from Amy's approval

* make codestyle

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

* update

* update

* update

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

* Experimental loading of MLX files

* Update exception message

* Add test

* Style

* Use model from hf-internal-testing

* post merge update

* nit

* oups

[tests] use the correct `n_gpu` in `TrainerIntegrationTest::test_train_and_eval_dataloaders` for XPU (#29307)

* fix n_gpu

* fix style

* Make sliding window size inclusive in eager attention

* Fix tests

* use torch_device

* skip for XPU

* Update tests/generation/test_utils.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

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

* fix image-to-text batch incorrect output issue
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>

* add ci test
Signed-off-by: Wang, Yi <yi.a.wang@intel.com>

* update ci test
Signed-off-by: Wang, Yi <yi.a.wang@intel.com>

---------
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
Signed-off-by: Wang, Yi <yi.a.wang@intel.com>

* add sacremoses check

* fix style

* for FlaubertTokenizer

* HerbertTokenizer fix

* add typeHint

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

* make less skipped

* make quality

* remove import

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

* left-padding test revisited

* Apply suggestions from code review
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

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

* 🐛 Fix vision encoder decoder positional arg

* ✅

 Add test for VisionEncoderDecoder with LayoutLMv3 encoder

---------
Co-authored-by: Nick DeGroot <1966472+nickthegroot@users.noreply.github.com>

* Fall back to pytorch model for now

* Fix up

* flava multimodal add attn mask

* make style

* check mask is not None

Revert "Automatic safetensors conversion when lacking these files (#29390)"

This reverts commit a69cbf4e.

* Enable BLIP for auto VQA

* Make style

* Add VQA to BLIP pipeline tests

* use require_torch_gpu

* enable on XPU

* fix

* Automatic safetensors conversion when lacking these files

* Remove debug

* Thread name

* Typo

* Ensure that raises do not affect the main thread

* initial-commit

* start cleaning

* small nits

* small nits

* current updates

* add kernels

* small refactoring little step

* add comments

* styling

* nit

* nits

* Style

* Small changes

* Push dummy mambda simple slow

* nit

* Use original names

* Use original names and remove norm

* Updates for inference params

* Style nd updates

* nits

* Match logits

* Add a test

* Add expected generated text

* nits doc, imports and styling

* style

* oups

* dont install kernels, invite users to install the required kernels

* let use use the original packages

* styling

* nits

* fix some copieds

* update doc

* fix-copies

* styling done

* nits

* fix import check

* run but wrong cuda ress

* mamba CUDA works :)

* fix the fast path

* config naming nits

* conversion script is not required at this stage

* finish fixing the fast path: generation make sense now!

* nit

* Let's start working on the CIs

* style

* better style

* more nits

* test nit

* quick fix for now

* nits

* nit

* nit

* nit

* nits

* update test rest

* fixup

* update test

* nit

* some fixes

* nits

* update test values

* fix styling

* nit

* support peft

* integrations tests require torchg

* also add slow markers

* styling

* chose forward wisely

* nits

* update tests

* fix gradient checkpointing

* fixup

* nit

* fix doc

* check copies

* fix the docstring

* fix some more tests

* style

* fix beam search

* add init schene

* update

* nit

* fix

* fixup the doc

* fix the doc

* fixup

* tentative update but slow is no longer good

* nit

* should we always use float32?

* nits

* revert wrong changes

* res in float32

* cleanup

* skip fmt for now

* update generation values

* update test values running original model

* fixup

* update tests + rename inference_params to cache_params + make sure training does not use cache_params

* small nits

* more nits

* fix final CIs

* style

* nit doc

* I hope final doc nits

* nit

* 🫠

* final touch!

* fix torch import

* Apply suggestions from code review
Co-authored-by: Lysandre Debut <hi@lysand.re>

* Apply suggestions from code review

* fix fix and fix

* fix base model prefix!

* nit

* Update src/transformers/models/mamba/__init__.py

* Update docs/source/en/model_doc/mamba.md
Co-authored-by: Lysandre Debut <hi@lysand.re>

* nit

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

* fix udop imports

* sort imports

* style

* revert with RP

* nit

* exact revert

* update

* ...

* nits

* arf

* 🧼

* beat the last guy

* style everyone

* use torch_device

* Update tests/pipelines/test_pipelines_text_generation.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* fix style

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

* added exllama kernels support for awq models

* doc

* style

* Update src/transformers/modeling_utils.py
Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>

* refactor

* moved exllama post init to after device dispatching

* bump autoawq version

* added exllama test

* style

* configurable exllama kernels

* copy exllama_config from gptq

* moved exllama version check to post init

* moved to quantization dockerfile

---------
Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>

* First draft

* More improvements

* More improvements

* More fixes

* Fix copies

* More improvements

* More fixes

* More improvements

* Convert checkpoint

* More improvements, set up tests

* Fix more tests

* Add UdopModel

* More improvements

* Fix equivalence test

* More fixes

* Redesign model

* Extend conversion script

* Use real inputs for conversion script

* Add image processor

* Improve conversion script

* Add UdopTokenizer

* Add fast tokenizer

* Add converter

* Update README's

* Add processor

* Add fully fledged tokenizer

* Add fast tokenizer

* Use processor in conversion script

* Add tokenizer tests

* Fix one more test

* Fix more tests

* Fix tokenizer tests

* Enable fast tokenizer tests

* Fix more tests

* Fix additional_special_tokens of fast tokenizer

* Fix tokenizer tests

* Fix more tests

* Fix equivalence test

* Rename image to pixel_values

* Rename seg_data to bbox

* More renamings

* Remove vis_special_token

* More improvements

* Add docs

* Fix copied from

* Update slow tokenizer

* Update fast tokenizer design

* Make text input optional

* Add first draft of processor tests

* Fix more processor tests

* Fix decoder_start_token_id

* Fix test_initialization

* Add integration test

* More improvements

* Improve processor, add test

* Add more copied from

* Add more copied from

* Add more copied from

* Add more copied from

* Remove print statement

* Update README and auto mapping

* Delete files

* Delete another file

* Remove code

* Fix test

* Fix docs

* Remove asserts

* Add doc tests

* Include UDOP in exotic model tests

* Add expected tesseract decodings

* Add sentencepiece

* Use same design as T5

* Add UdopEncoderModel

* Add UdopEncoderModel to tests

* More fixes

* Fix fast tokenizer

* Fix one more test

* Remove parallelisable attribute

* Fix copies

* Remove legacy file

* Copy from T5Tokenizer

* Fix rebase

* More fixes, copy from T5

* More fixes

* Fix init

* Use ArthurZ/udop for tests

* Make all model tests pass

* Remove UdopForConditionalGeneration from auto mapping

* Fix more tests

* fixups

* more fixups

* fix the tokenizers

* remove un-necessary changes

* nits

* nits

* replace truncate_sequences_boxes with truncate_sequences for fix-copies

* nit current path

* add a test for input ids

* ids that we should get taken from c9f7a32f57440d90ff79890270d376a1cc0acb68

* nits converting

* nits

* apply ruff

* nits

* nits

* style

* fix slow order of addition

* fix udop fast range as well

* fixup

* nits

* Add docstrings

* Fix gradient checkpointing

* Update code examples

* Skip tests

* Update integration test

* Address comment

* Make fixup

* Remove extra ids from tokenizer

* Skip test

* Apply suggestions from code review
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* Update year

* Address comment

* Address more comments

* Address comments

* Add copied from

* Update CI

* Rename script

* Update model id

* Add AddedToken, skip tests

* Update CI

* Fix doc tests

* Do not use Tesseract for the doc tests

* Remove kwargs

* Add original inputs

* Update casting

* Fix doc test

* Update question

* Update question

* Use LayoutLMv3ImageProcessor

* Update organization

* Improve docs

* Update forward signature

* Make images optional

* Remove deprecated device argument

* Add comment, add add_prefix_space

* More improvements

* Remove kwargs

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

* Update ms_deform_attn_cuda.cu

* Update ms_deform_attn_cuda.cuh

* Update modeling_deformable_detr.py

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

* Update modeling_deformable_detr.py

* python utils/check_copies.py --fix_and_overwrite

* Fix dtype missmatch error

* Update test_modeling_deformable_detr.py

* Update test_modeling_deformable_detr.py

* Update modeling_deformable_detr.py

* Update modeling_deformable_detr.py

* Support DeformableDETR with bfloat16

* Add test code

* Use AT_DISPATCH_FLOATING_TYPES_AND2

Use AT_DISPATCH_FLOATING_TYPES_AND2

* Update tests/models/deformable_detr/test_modeling_deformable_detr.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Update tests/models/deformable_detr/test_modeling_deformable_detr.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Fix not found require_torch_bf16 function

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

Fully revert atomic checkpointing

* 🐛 Fix oneformer instance post processing when using panoptic task type

* ✅

 Add unit test for oneformer instance post processing panoptic bug

---------
Co-authored-by: Nick DeGroot <1966472+nickthegroot@users.noreply.github.com>

* use require_torch_gpu

* enable on XPU

* Fix deprecated arg issue

* Trainer check too

* Check for dict or dataclass

* Simplify, make config always AcceleratorConfig

* Upstream to Trainer

* Fix yolos processing

* Add back slow marker - protects for pycocotools in slow

* Slow decorator goes above copied from header

* [Whisper Tok] Update integration test

* make style