GitLab

LayoutLMv3TokenizerFast produces empty 'Ġ' token with `offset_mapping = (0, 0)`.
Next token is wrongly assumed to also be beginning of word and isn't
correctly assigned `pad_token_label`.
Modify test with text that produce 'Ġ' token.
Remove copy check from LayoutLMv2TokenizerFast for `_batch_encode_plus`.

solves issue: #19978

* haha text go brrr (but in gradio)

* added biogpt token classifier

* fix reviews

* Updated modeling_biogpt.py
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>

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

* draft

* update tokenization limma and conversion script

* more udpates

* initial commit

* style

* default pad to None

* draft tokenization tests

* update test

* update tokenization tests

* nits

* update

* versioning test

* major fix

* fix more testst

* finish fixing special masks

* last nit

* more nits

* add encode decode tests

* add more

* fix token type ids

* style

added > 0.5 to `past_observed_mask`

* Test fetcher v2

* Fix regexes

* Remove sanity check

* Fake modification to OPT

* Fixes some .sep issues

* Remove fake OPT change

* Fake modif for BERT

* Fake modif for init

* Exclude SageMaker tests

* Fix test and remove fake modif

* Fake setup modif

* Fake pipeline modif

* Remove all fake modifs

* Adds options to skip/force tests

* [test-all-models] Fake modif for BERT

* Try this way

* Does the command actually work?

* [test-all-models] Try again!

* [skip circleci] Remove fake modif

* Remove debug statements

* Add the list of important models

* Quality

* Update utils/tests_fetcher.py
Co-authored-by: Lysandre Debut <lysandre.debut@reseau.eseo.fr>

* Address review comments

* Address review comments

* Fix and add test

* Apply suggestions from code review
Co-authored-by: Yih-Dar <2521628+ydshieh@users.noreply.github.com>

* Address review comments

---------
Co-authored-by: Lysandre Debut <lysandre.debut@reseau.eseo.fr>
Co-authored-by: Yih-Dar <2521628+ydshieh@users.noreply.github.com>

* Making sure we can use safetensors to serialize all the time.

* Expanding the tests for increased coverage.

* Update the test.

* Getting current state of affairs.

* Tentative fix.

* Fixing black version.

* Fixing the worst offenders.

* Try to modify less files.

* Fixing blip_2 (Weird solution right now).

* Fixing deta.

* Fix blip ?

* Missing extra newline.

* No deta modification.

* Adding some comments.

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

* Addressing comments.

* Addressing comments.

* creating warn_once.

* Warning_once !

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

edit default model type and testing path set to hf-internal-testing

* haha tokens go brrrr

Skip flaky test for now

* Rescale image back if it was scaled during PIL conversion

* do_rescale is defined if PIL image passed in

fix slow test

Fix some tiny model creation issues
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* fix bnb failing test

* fix

* fix

* fixup

* add draft changes

* fix failing wav2vec

* style

* make sure that the argument is saved + add tests

* style

* fixup

* update test

* default clean_up_tokenization_spaces to False for Bloom and Llama

* Update code based on review
Co-authored-by: Nicolas Patry <patry.nicolas@gmail.com>

* style

* quality

---------
Co-authored-by: Nicolas Patry <patry.nicolas@gmail.com>

* Initial commit

* update modeling code

* update doc

* add functions necessary

* fix impotrs

* revert changes

* fixup

* more styling to get going

* remove standalone encoder

* update code

* styling

* fix config and model

* update code and some refactoring

* make more tests pass

* Adding NLLB-200 - MoE - 54.5B for no language left behind
Fixes #21300

* fix mor common tests

* styke

* update testing file

* update

* update

* Router2 doc

* update check config with sparse layer

* add dummy router

* update current conversion script

* create on the fly conversion script

* Fixup

* style

* style 2

* fix empty return

* fix return

* Update default config sparse layers

* easier to create sparse layers

* update

* update conversion script

* update modeling

* add to toctree

* styling

* make ruff happy

* update docstring

* update conversion script

* update, will break tests but impelemting top2

* update

* ❗local groups are supported here

* ⚠️ Support for local groups is now removed ⚠️

This is because it has to work with model parallelism that we do not support

* finish simplificaiton

* Fix forward

* style

* fixup

* Update modelling and test, refactoring

* update tests

* remove final layer)norm as it is done in the FF

* routing works! Logits test added

* nit in test

* remove top1router

* style

* make sure sparse are tested. Had to change route_tokens a liottle bit

* add support for unslip models when converting

* fixup

* style

* update test s

* update test

* REFACTOR

* encoder outputs match!

* style

* update testing

* 🎉encoder and decoder logits match 🎉



* styleing

* update tests

* cleanup tests

* fix router test and CIs

* cleanup

* cleanup test styling

* fix tests

* Finally the generation tests match!

* cleanup

* update test

* style testing file

* remove script

* cleanup

* more cleanup

* nits

* update

* NLLB tokenizer is wrong and will be fixed soon

* use LongTensors

* update tests

* revert some small changes

* fix second expert sampling and batch prioritized routing

* update tests

* finish last tests

* make ruff happy

* update

* ruff again

* style

* Update docs/source/en/model_doc/nllb-moe.mdx
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Updates based on review

* style and fix import issue

* nit

* more nits

* cleanup

* styling

* update test_seconde_expert_policy

* fix name

* last nit on the markdown examples

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

* First draft

* Fix integration test

* Remove script

* Fix test and typos

* Fix one more test

* Skip tied embeddings test

* Remove line

* Address comments

* [WIP] flax resnet

* added pretrained flax models, results reproducible

* Added pretrained flax models, results reproducible

* working on tests

* no real code change, just some comments

* [flax] adding support for batch norm layers

* fixing bugs related to pt+flax integration

* removing loss from modeling flax output class

* fixing classifier tests

* fixing comments, model output

* cleaning comments

* review changes

* review changes

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

* renaming Flax to PyTorch

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

* add mega file structure and plain pytorch version of mega source code

* added config class with old naming conventions

* filled in mega documentation

* added config class and embeddings with optional token types

* updated notes

* starting the conversion process, deleted intermediate and added use_cache back to config

* renamed config attributes in modeling_mega.py

* checkpointing before refactoring incremental decoding functions

* removed stateful incremental key/values for EMA and self-attention

* refactored MovingAverageGatedAttention to remove stateful k/v history and use unified attention mask

* MovingAverageGatedAttention works with incremental decoding + past values, added sequence length enforcement

* more comments in MovingAverageGatedAttention + checkpointing before GatedCrossAttention

* bug fix in attention mask handling in MovingAverageGatedAttention

* removed incremental state from GatedCrossAttention and removed IncrementalState class

* finished gated cross attention and got MegaLayer working

* fixed causal masking in mega decoder

* fixed how padding and causal masks are passed through MegaLayer with and without k/v caching

* finished MegaModel; tested with encoder, decoder-only, and cross-attention type inputs; started work on downstream classes; removed mentions of position_ids

* added optional dense hidden layer for masked and causal LM classes

* docstring updates in MultiHeadEMA and GatedCrossAttention, removed unnecessary inputs in cross-attention

* removed before_attn_fn in Mega class and updated docstrings and comments up to there

* bug fix in MovingAverageGatedAttention masking

* working conversion of MLM checkpoint in scratchpad script -- perfect matches

* moved arg for hidden dense layer in LM head to config; discovered issue where from_pretrained is renaming gamma and beta parameters

* renamed gamma and beta parameters to avoid HF renaming when loading from checkpoint

* finished checkpoint conversion script

* cleanup old class in mega config script

* removed 'copied from' statements and passing integration tests

* added num_attention_heads=1 to config for integration compatibility, decoder tests working, generation tests failing

* fixed tuple output of megamodel

* all common tests passing after fixing issues in decoder, gradient retention, and initialization

* added mega-specific tests, ready for more documentation and style checks

* updated docstrings; checkpoint before style fixes

* style and quality checks, fixed initialization problem in float_tensor, ready for PR

* added mega to toctree

* removed unnecessary arg in megaconfig

* removed unused arg and fixed code samples with leftover roberta models

* Apply suggestions from code review

Applied all suggestions except the one renaming a class, as I'll need to update that througout
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* fixed issue where .view breaks batch dimension, conversion script fixed with absolute imports, updated readme with Mega->MEGA

* removed asserts in Mega code, renamed sequencenorm, gatedcrossattention, and NFFN, replaced get_activation_fn with ACTFN, and added sequencenorm to layer norms

* reformatted .forward() docstrings to match style and removed unused mask input in cross-attention

* removed all reset_parameters() methods and rolled into MegaPreTrainedModel._init_weights()

* renamed all single-letter variables and improved readability in tensor size comments, Mega->MEGA in 2 documentation files

* variable names in NFFN

* manual Mega->MEGA changes in docs

* Mega->MEGA in config auto

* style and quality fixes

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

* renamed parameters and variables with confusing names, added copied from statements, moved fft conv to its own method, other cleanup from PR comments

* commit before dealing with merge conflicts

* made new attention activation functions available in ACT2FN and added generation test from OPT

* style and quality in activations and tests

* documentation fixes, renaming variables in dropout and rotary positions, used built-in causal masking, encoders->layers in MegaModel, moved comments into docstrings

* style and quality fixes after latest updates, before rotary position ids

* causal mask in MegaBlock docstring + added missing device passing

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

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

* added Mega prefixes where missing, reverted MegaSequenceNorm to if-else, other module renaming requested in PR

* style and quality fixes + readme updates pointing to main

---------
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Automatically create or update tiny models

* Skip failed tests

* update workflow file

* use revision

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

* check what tests fail

* Skip failing tests

* Skip failing tests

* Skip failing tests

* Skip failing tests

* clean up

* clean up

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

* Chunkable classification pipeline 

The TokenClassificationPipeline is now able to process sequences longer than 512. No matter the framework, the model, the tokenizer. We just have to pass process_all=True and a stride number (optional). The behavior remains the same if you don't pass these optional parameters. For overlapping parts when using stride above 0, we consider only the max scores for each overlapped token in all chunks where the token is.

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* update with latest black format

* update black format

* Update token_classification.py

* Update token_classification.py

* format correction

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update comments

* Update src/transformers/pipelines/token_classification.py
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>

* Update token_classification.py

Correct spaces, remove process_all and keep only stride. If stride is provided, the pipeline is applied to the whole text.

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update chunk aggregation

Update the chunk aggregation strategy based on entities aggregation.

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

Remove unnecessary pop from outputs dict

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

* Update token_classification.py

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

* add chunking tests

* correct formating

* correct formatting

* correct model id for test chunking

* update scores with nested simplify

* Update test_pipelines_token_classification.py

* Update test_pipelines_token_classification.py

* update model to a tiny one

* Update test_pipelines_token_classification.py

* Adding smaller test for chunking.

* Fixup

* Update token_classification.py

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

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

---------
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* v1 all keys match

* clean up

* forward pass ok

* add correct image transform

* generate works, logits matching

* clean up

* more refactor

* revert

* revert

* clean up

* clean ups

* clean up

* refactor

* refactor

* fix doc

* fix tokenizer test

* fix toctree

* revert toctree

* oops

* few fixes

* replace to `pixel_embeds`

* make fixup

* test processing & feat extractor

* fix some tests

* more fixes

* make fixup

* clean up

* more clean up

* add a single slow test

* fix test

* make fixup

* fix

* fix authors

* fix toctree

* update docs

* add docstring

* revert change

* Update src/transformers/models/pix2struct/__init__.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* fix tokenizer

* fix processor test

* fix test

* make fixup

* refactor

* fix config

* Update src/transformers/models/pix2struct/image_processing_pix2struct.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* format

* fix

* Update src/transformers/models/pix2struct/image_processing_pix2struct.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* make fixup

* add docstring

* fix issues

* fix

* fix

* fix

* add slow test

* fix

* fix

* fix batched issue

* fix training issues

* fix ci test

* fix slow test

* fix conversion script

* remove unneeded classes

* fix slow test

* fix require backends

* fix masked fill

* revert

* fix softmax

* add large models support

* fix conditional generation

* few fixes

* add instructions

* rm unneeded file

* Update src/transformers/models/pix2struct/convert_pix2struct_original_pytorch_to_hf.py

* fix ci test

* fix ci test really

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

* fix nit

* fix nits

* fix image processors nits

* docstring

* clean up

* fix nit

* fix tests

* docstring nit

* fix reshape

* Update src/transformers/models/pix2struct/image_processing_pix2struct.py
Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>

* fix nit

* fix repetition

* refactor processor

* make patch size consistent

* refactor forward

* fix docstring

* fix max_patches issue

* update docstirng

* update docstring

* fix coped from

* add skip reasons

* few fixes

* Update src/transformers/models/pix2struct/image_processing_pix2struct.py
Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>

* format

* fix doctests

* refactor and fix

* fix doc build issue

* fix processor test

* small fix conversion script

* replace correct weights

* make fixup

* fix some issues

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

* revert config and fixes

* Update src/transformers/models/pix2struct/image_processing_pix2struct.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* more details

* fixes

* fix processor

* fix processor test

* fix

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

* make fixup

* fix processor

* Update src/transformers/models/pix2struct/modeling_pix2struct.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* add copied

* make fixup

* fix copies

* update docstring

* refactor

* fix docstring

* fix conversion script

* fix vqa issue

* replace to `flattened_patches`

* nit

* fix numpy issue

* fix image processors

* add batched vqa support

* fix vqa conversion

* make fixup

* fix conversion script

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

* make fixup

* add correct docstring

* update docstring

* fix module level + channel dim

* use `make_list_of_images`

* refactor

* correct docstring

* fix authors

* remove `data_format`

* add header text test

* Apply suggestions from code review
Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* make fixup

* add checkpoints

---------
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>

* tmp commit

* beef up llama tests

* Export TF generate with a TF tokenizer

* remove unused lines

Fixed bug to calculate correct xpath_sub_list in MarkupLMTokenizer. Earlier xpath_sub_list was same as xpath_tags_list
Co-authored-by: dusejat <dusejat@amazon.com>

* Add MaskedImageModelingOutput

* time to say goodbye, torch 1.7 and 1.8

* clean up torch_int_div

* clean up is_torch_less_than_1_8-9

* update

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

* Make sure CVT can be trained using mixed precision

* Add test for keras-fit with mixed-precision

* Update tests/models/cvt/test_modeling_tf_cvt.py
Co-authored-by: Matt <Rocketknight1@users.noreply.github.com>

---------
Co-authored-by: gcuder <Gerald.Cuder@iacapps.com>
Co-authored-by: Matt <Rocketknight1@users.noreply.github.com>

* Add LlamaForSequenceClassification

* Update src/transformers/models/llama/modeling_llama.py
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>

* Update src/transformers/models/llama/modeling_llama.py
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>

* Add docstring

* Add test

* Add input embedding getter and setter

* Remove dead code

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

* py38 + torch 2

* increment cache versions

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

* LLaMA

* sharding and docs

* tweak

* black

* inits

* ruff

* LLAMA_PRETRAINED_CONFIG_ARCHIVE_MAP

* init

* no checkpoint

* docs

* ruff

* type_vocab_size

* tokenizer fixes

* tokenizer fixes

* Update tokenization_llama.py

* Update tokenization_llama.py

* Update configuration_llama.py

* Update modeling_llama.py

* tokenizer add_bos by default

* licenses

* remove decoder

* norms and mlp

* rope overhaul

* tweaks

* black

* mention OPT implementation

* off-by-one naming

* typo

* fix

* tokenization fix and slicing bug

* padding config

* cleanup

* black

* update tests

* undo typo

* fix vocab caching logic

* ruff

* docbuilder

* attn fix from BlackSamorez

* initial feedback

* typo

* docs

* llama case

* llama case

* load checkpoint docs

* comment about tokenizer

* tokenizer defaults

* clear past_key_values if use_cache=False

* last tweaks

* last tweaks

* last tweaks

* last tweaks

---------
Co-authored-by: Stella Biderman <stellabiderman@gmail.com>

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

* Use return_loss for BridgeTowerForContrastiveLearning, add example

* fix tests

* Update example in BridgeTowerForContrastiveLearning

* Update test_modeling_bridgetower.py

* update model output format

* minor update

* Update src/transformers/models/bridgetower/modeling_bridgetower.py

* make style

---------
Co-authored-by: Tiep Le <97980157+tileintel@users.noreply.github.com>
Co-authored-by: Tiep Le <tiep.le@intel.com>
Co-authored-by: Yih-Dar <2521628+ydshieh@users.noreply.github.com>
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* Fix regression in pipeline when device=-1 is passed

* Add regression test