* make the save_load special key tests common
* handle mbart
* cleaner solution
* fix
* move test_save_load_missing_keys back into fstm for now
* restore
* style
* add marian
* add pegasus
* blenderbot
* revert - no static embed
* add CustomHFIndex
* typo in config
* update tests
* add custom dataset example
* clean script
* update test data
* minor in test
* docs
* docs
* style
* fix imports
* allow to pass the indexed dataset directly
* update tests
* use multiset DPR
* address thom and patrick's comments
* style
* update dpr tokenizer
* add output_dir flag in use_own_knowledge_dataset.py
* allow custom datasets in examples/rag/finetune.py
* add test for custom dataset in distributed rag retriever
* fix 5990
* accomodate iterable dataset without predefined length
* set it as 1 use case: provide max_steps, and NO num_epochs
* Is a merge of master and PR 5995
* fix trainer test under TF
* fix only for torch
* TF trainer untouched
* trainer tests are skipped when no torch
* address comments
* fix quality checks
* remove torch.dataset from test_trainer
* unnecessary inheritance
* RegressionDataset implements all needed methods __len__ and __getitem__
* fix quality checks
* restore RegressionDataset
* was wrongly under is_torch_available()
* WIP flax bert
* Initial commit Bert Jax/Flax implementation.
* Embeddings working and equivalent to PyTorch.
* Move embeddings in its own module BertEmbeddings
* Added jax.jit annotation on forward call
* BertEncoder on par with PyTorch ! :D
* Add BertPooler on par with PyTorch !!
* Working Jax+Flax implementation of BertModel with < 1e-5 differences on the last layer.
* Fix pooled output to take only the first token of the sequence.
* Refactoring to use BertConfig from transformers.
* Renamed FXBertModel to FlaxBertModel
* Model is now initialized in FlaxBertModel constructor and reused.
* WIP JaxPreTrainedModel
* Cleaning up the code of FlaxBertModel
* Added ability to load Flax model saved through save_pretrained()
* Added ability to convert Pytorch Bert model to FlaxBert
* FlaxBert can now load every Pytorch Bert model with on-the-fly conversion
* Fix hardcoded shape values in conversion scripts.
* Improve the way we handle LayerNorm conversion from PyTorch to Flax.
* Added positional embeddings as parameter of BertModel with default to np.arange.
* Let's roll FlaxRoberta !
* Fix missing position_ids parameters on predict for Bert
* Flax backend now supports batched inputs
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>
* Make it possible to load msgpacked model on convert from pytorch in last resort.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>
* Moved save_pretrained to Jax base class along with more constructor parameters.
* Use specialized, model dependent conversion functio.
* Expose `is_flax_available` in file_utils.
* Added unittest for Flax models.
* Added run_tests_flax to the CI.
* Introduce FlaxAutoModel
* Added more unittests
* Flax model reference the _MODEL_ARCHIVE_MAP from PyTorch model.
* Addressing review comments.
* Expose seed in both Bert and Roberta
* Fix typo suggested by @stefan-it
Co-Authored-By: Stefan Schweter <stefan@schweter.it>
* Attempt to make style
* Attempt to make style in tests too
* Added jax & jaxlib to the flax optional dependencies.
* Attempt to fix flake8 warnings ...
* Redo black again and again
* When black and flake8 fight each other for a space ... 💥💥💥
* Try removing trailing comma to make both black and flake happy!
* Fix invalid is_<framework>_available call, thanks @LysandreJik 🎉
* Fix another invalid import in flax_roberta test
* Bump and pin flax release to 0.1.0.
* Make flake8 happy, remove unused jax import
* Change the type of the catch for msgpack.
* Remove unused import.
* Put seed as optional constructor parameter.
* trigger ci again
* Fix too much parameters in BertAttention.
* Formatting.
* Simplify Flax unittests to avoid machine crashes.
* Fix invalid number of arguments when raising issue for an unknown model.
* Address @bastings comment in PR, moving jax.jit decorated outside of __call__
* Fix incorrect path to require_flax/require_pytorch functions.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Attempt to make style.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Correct rebasing of circle-ci dependencies
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Fix import sorting.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Fix unused imports.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Again import sorting...
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Installing missing nlp dependency for flax unittests.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Fix laoding of model for Flax implementations.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* jit the inner function call to make JAX-compatible
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Format !
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Flake one more time 🎶
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Rewrites BERT in Flax to the new Linen API (#7211)
* Rewrite Flax HuggingFace PR to Linen
* Some fixes
* Fix tests
* Fix CI with change of name of nlp (#7054)
* nlp -> datasets
* More nlp -> datasets
* Woopsie
* More nlp -> datasets
* One last
* Expose `is_flax_available` in file_utils.
* Added run_tests_flax to the CI.
* Attempt to make style
* trigger ci again
* Fix import sorting.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Revert "Rewrites BERT in Flax to the new Linen API (#7211)"
This reverts commit 23703a5eb3364e26a1cbc3ee34b4710d86a674b0.
* Remove jnp.lax references
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Make style.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Reintroduce Linen changes ...
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Make style.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Use jax native's gelu function.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Renaming BertModel to BertModule to highlight the fact this is the Flax Module object.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Rewrite FlaxAutoModel test to not rely on pretrained_model_archive_map
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Remove unused variable in BertModule.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Remove unused variable in BertModule again
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Attempt to have is_flax_available working again.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Introduce JAX TensorType
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>
* Improve ImportError message when trying to convert to various TensorType format.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>
* Makes Flax model jittable.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>
* Ensure flax models are jittable in unittests.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>
* Remove unused imports.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Ensure jax imports are guarded behind is_flax_available.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Make style.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Make style again
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Make style again again
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Make style again again again
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Update src/transformers/file_utils.py
Co-authored-by: Marc van Zee <marcvanzee@gmail.com>
* Bump flax to it's latest version
Co-authored-by: Marc van Zee <marcvanzee@gmail.com>
* Bump jax version to at least 0.2.0
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Style.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Update the unittest to use TensorType.JAX
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* isort import in tests.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Match new flax parameters name "params"
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Remove unused imports.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Add flax models to transformers __init__
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Attempt to address all CI related comments.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Correct circle.yml indent.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Correct circle.yml indent (2)
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Remove coverage from flax tests
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Addressing many naming suggestions from comments
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Simplify for loop logic to interate over layers in FlaxBertLayerCollection
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* use f-string syntax for formatting logs.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Use config property from FlaxPreTrainedModel.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* use "cls_token" instead of "first_token" variable name.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* use "hidden_state" instead of "h" variable name.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Correct class reference in docstring to link to Flax related modules.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Added HF + Google Flax team copyright.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Make Roberta independent from Bert
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Move activation functions to flax_utils.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Move activation functions to flax_utils for bert.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Added docstring for BERT
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Update import for Bert and Roberta tokenizers
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Make style.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* fix-copies
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Correct FlaxRobertaLayer to match PyTorch.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Use the same store_artifact for flax unittest
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Style.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Make sure gradient are disabled only locally for flax unittest using torch equivalence.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Use relative imports
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
Co-authored-by: Stefan Schweter <stefan@schweter.it>
Co-authored-by: Marc van Zee <marcvanzee@gmail.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
* Propagating n_docs as parameter to all RagModel's related functions that defaults to self.config.n_docs
* Making n_docs parameter's default value to None in marginalize function
* Fixing code quality issues
* Handle the special case when generator is of T5PreTrainedModel instance type. T5PreTrainedModel do not have n_docs as parameter
* T5PreTrainedModel do not have n_docs as parameter
* Addressing review comment
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* Correcting comment by addressing review comment
* Adding assert statement verifying that n_docs is correctly set. n_docs should be the same for both retriever and generator.
* Fixing flake8 reported issue
* Correcting test datasets for rag
* Using doc_scores instead of context_input_ids to check assert as in RagSequenceForGeneration context_input_ids can be null
* doc_scores second dimension have number of retrieved docs
* Changing assert comment
* Apply suggestions from code review
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* splitting fast and slow tokenizers [WIP]
* [WIP] splitting sentencepiece and tokenizers dependencies
* update dummy objects
* add name_or_path to models and tokenizers
* prefix added to file names
* prefix
* styling + quality
* spliting all the tokenizer files - sorting sentencepiece based ones
* update tokenizer version up to 0.9.0
* remove hard dependency on sentencepiece 🎉
* and removed hard dependency on tokenizers 🎉
* update conversion script
* update missing models
* fixing tests
* move test_tokenization_fast to main tokenization tests - fix bugs
* bump up tokenizers
* fix bert_generation
* update ad fix several tokenizers
* keep sentencepiece in deps for now
* fix funnel and deberta tests
* fix fsmt
* fix marian tests
* fix layoutlm
* fix squeezebert and gpt2
* fix T5 tokenization
* fix xlnet tests
* style
* fix mbart
* bump up tokenizers to 0.9.2
* fix model tests
* fix tf models
* fix seq2seq examples
* fix tests without sentencepiece
* fix slow => fast conversion without sentencepiece
* update auto and bert generation tests
* fix mbart tests
* fix auto and common test without tokenizers
* fix tests without tokenizers
* clean up tests lighten up when tokenizers + sentencepiece are both off
* style quality and tests fixing
* add sentencepiece to doc/examples reqs
* leave sentencepiece on for now
* style quality split hebert and fix pegasus
* WIP Herbert fast
* add sample_text_no_unicode and fix hebert tokenization
* skip FSMT example test for now
* fix style
* fix fsmt in example tests
* update following Lysandre and Sylvain's comments
* Update src/transformers/testing_utils.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
* Update src/transformers/testing_utils.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
* Update src/transformers/tokenization_utils_base.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
* Update src/transformers/tokenization_utils_base.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
* HerBERT transformer model for Polish language understanding.
* HerbertTokenizerFast generated with HerbertConverter
* Herbert base and large model cards
* Herbert model cards with tags
* Herbert tensorflow models
* Herbert model tests based on Bert test suit
* src/transformers/tokenization_herbert.py edited online with Bitbucket
* src/transformers/tokenization_herbert.py edited online with Bitbucket
* docs/source/model_doc/herbert.rst edited online with Bitbucket
* Herbert tokenizer tests and bug fixes
* src/transformers/configuration_herbert.py edited online with Bitbucket
* Copyrights and tests for TFHerbertModel
* model_cards/allegro/herbert-base-cased/README.md edited online with Bitbucket
* model_cards/allegro/herbert-large-cased/README.md edited online with Bitbucket
* Bug fixes after testing
* Reformat modified_only_fixup
* Proper order of configuration
* Herbert proper documentation formatting
* Formatting with make modified_only_fixup
* Dummies fixed
* Adding missing models to documentation
* Removing HerBERT model as it is a simple extension of BERT
* Update model_cards/allegro/herbert-base-cased/README.md
Co-authored-by: Julien Chaumond <chaumond@gmail.com>
* Update model_cards/allegro/herbert-large-cased/README.md
Co-authored-by: Julien Chaumond <chaumond@gmail.com>
* HerbertTokenizer deprecated configuration removed
Co-authored-by: Julien Chaumond <chaumond@gmail.com>
* Improving Pipelines by defaulting to framework='tf' when
pytorch seems unavailable.
* Actually changing the default resolution order to account for model
defaults
Adding a new tests for each pipeline to check that pipeline(task) works
too without manually adding the framework too.
* Add Documentation for GPT-1 Classification
* Add GPT-1 with Classification head
* Add tests for GPT-1 Classification
* Add GPT-1 For Classification to auto models
* Remove authorized missing keys, change checkpoint to openai-gpt
* Reintroduce clean_text call which was removed by mistake in #4723
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Added unittest for clean_text parameter on Bert tokenizer.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Better unittest name.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Adapt unittest to use untrained tokenizer.
Signed-off-by: Morgan Funtowicz <funtowiczmo@gmail.com>
* Code quality + update test
Co-authored-by: Lysandre <lysandre.debut@reseau.eseo.fr>
* [WIP] SP tokenizers
* fixing tests for T5
* WIP tokenizers
* serialization
* update T5
* WIP T5 tokenization
* slow to fast conversion script
* Refactoring to move tokenzier implementations inside transformers
* Adding gpt - refactoring - quality
* WIP adding several tokenizers to the fast world
* WIP Roberta - moving implementations
* update to dev4 switch file loading to in-memory loading
* Updating and fixing
* advancing on the tokenizers - updating do_lower_case
* style and quality
* moving forward with tokenizers conversion and tests
* MBart, T5
* dumping the fast version of transformer XL
* Adding to autotokenizers + style/quality
* update init and space_between_special_tokens
* style and quality
* bump up tokenizers version
* add protobuf
* fix pickle Bert JP with Mecab
* fix newly added tokenizers
* style and quality
* fix bert japanese
* fix funnel
* limite tokenizer warning to one occurence
* clean up file
* fix new tokenizers
* fast tokenizers deep tests
* WIP adding all the special fast tests on the new fast tokenizers
* quick fix
* adding more fast tokenizers in the fast tests
* all tokenizers in fast version tested
* Adding BertGenerationFast
* bump up setup.py for CI
* remove BertGenerationFast (too early)
* bump up tokenizers version
* Clean old docstrings
* Typo
* Update following Lysandre comments
Co-authored-by: Sylvain Gugger <sylvain.gugger@gmail.com>
* Initial callback proposal
* Finish various callbacks
* Post-rebase conflicts
* Fix tests
* Don't use something that's not set
* Documentation
* Remove unwanted print.
* Document all models can work
* Add tests + small fixes
* Update docs/source/internal/trainer_utils.rst
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>
* Address review comments
* Fix TF tests
* Real fix this time
* This one should work
* Fix typo
* Really fix typo
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>
* configuration_squeezebert.py
thin wrapper around bert tokenizer
fix typos
wip sb model code
wip modeling_squeezebert.py. Next step is to get the multi-layer-output interface working
set up squeezebert to use BertModelOutput when returning results.
squeezebert documentation
formatting
allow head mask that is an array of [None, ..., None]
docs
docs cont'd
path to vocab
docs and pointers to cloud files (WIP)
line length and indentation
squeezebert model cards
formatting of model cards
untrack modeling_squeezebert_scratchpad.py
update aws paths to vocab and config files
get rid of stub of NSP code, and advise users to pretrain with mlm only
fix rebase issues
redo rebase of modeling_auto.py
fix issues with code formatting
more code format auto-fixes
move squeezebert before bert in tokenization_auto.py and modeling_auto.py because squeezebert inherits from bert
tests for squeezebert modeling and tokenization
fix typo
move squeezebert before bert in modeling_auto.py to fix inheritance problem
disable test_head_masking, since squeezebert doesn't yet implement head masking
fix issues exposed by the test_modeling_squeezebert.py
fix an issue exposed by test_tokenization_squeezebert.py
fix issue exposed by test_modeling_squeezebert.py
auto generated code style improvement
issue that we inherited from modeling_xxx.py: SqueezeBertForMaskedLM.forward() calls self.cls(), but there is no self.cls, and I think the goal was actually to call self.lm_head()
update copyright
resolve failing 'test_hidden_states_output' and remove unused encoder_hidden_states and encoder_attention_mask
docs
add integration test. rename squeezebert-mnli --> squeezebert/squeezebert-mnli
autogenerated formatting tweaks
integrate feedback from patrickvonplaten and sgugger to programming style and documentation strings
* tiny change to order of imports
* Trainer should not modify its TrainingArguments
* Trainer should not modify its TrainingArguments
* Trainer should not modify its TrainingArguments
* Add test of resumed training
* Fixes
* Non multiGPU test
* Clean Trainer state
* Add more to the state
* Documentation
* One last test
* Make resume training test more complete
* Unwanted changes
* GPT2 gradient checkpointing
* find_unused_parameters removed if checkpointing
* find_unused_parameters removed if checkpointing
* Update src/transformers/configuration_gpt2.py
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* Added a test for generation with checkpointing
* Update src/transformers/configuration_gpt2.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
