GitLab

* add first files

* add xlm roberta integration tests

* make style

* flake 8 issues solved

* change masking to direct labelings

* fix black

* switch to ignore index

* .

* fix black

* Usage: Sequence Classification & Question Answering

* Pipeline example

* Language modeling

* TensorFlow code for Sequence classification

* Custom TF/PT toggler in docs

* QA + LM for TensorFlow

* Finish Usage for both PyTorch and TensorFlow

* Addressing Julien's comments

* More assertive

* cleanup

* Favicon
- added favicon option in conf.py along with the favicon image
- udpated 🤗

 logo. slightly smaller and should appear more consistent across editing programs (no more tongue on the outside of the mouth)
Co-authored-by: joshchagani <joshua@joshuachagani.com>

* Renamed file generate by tokenizers when calling save_pretrained to match python.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* Added save_vocabulary tests.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* Remove python quick and dirty fix for clean Rust impl.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* Bump tokenizers dependency to 0.5.1
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* TransfoXLTokenizerFast uses a json vocabulary file + warning about incompatibility between Python and Rust
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* Added some save_pretrained / from_pretrained unittests.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* Update tokenizers to 0.5.2
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* Quality and format.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* flake8
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* Making sure there is really a bug in unittest

* Fix TransfoXL constructor vocab_file / pretrained_vocab_file mixin.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* add explaining example to XLNet LM modeling

* improve docstring for xlnet

Add preprocessing step for transfo-xl tokenization to avoid tokenizing words followed by punction to <unk> (#2987)

* add preprocessing to add space before punctuation for transfo_xl

* improve warning messages

* make style

* compile regex at instantination of tokenizer object

* Add disable_outgoing to pretrained items

Setting disable_outgoing=True disables outgonig traffic:
- etags are not looked up
- models are not downloaded

* parameter name change

* Remove forgotten print

* Testing that encode_plus and batch_encode_plus behave the same way

Spoiler alert: they don't

* Testing rest of arguments in batch_encode_plus

* Test tensor return in batch_encode_plus

* Addressing Sam's comments

* flake8

* Simplified with `num_added_tokens`

* add slow generate lm_model tests

* fix conflicts

* merge conflicts

* fix conflicts

* add slow generate lm_model tests

* make style

* delete unused variable

* fix conflicts

* fix conflicts

* fix conflicts

* delete unused variable

* fix conflicts

* finished hard coded tests

Warning on `add_special_tokens` when passed to `encode`, `encode_plus` and `batch_encode_plus`

Don't know of a use case where that would be useful, but this is more consistent

Add image

- I added an example using the model with pipelines to show that we have set```{"use_fast": False}``` in the tokenizer.
- I added a Colab to play with the model and pipelines
- I added a Colab to discover Huggingface pipelines at the end of the document

* enable_padding should pad up to max_length if set.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* Added more testing on padding.
Signed-off-by: Morgan Funtowicz <morgan@huggingface.co>

* Only use F.gelu for torch >=1.4.0

* Use F.gelu for newer torch

* improving generation

* finalized special token behaviour for no_beam_search generation

* solved modeling_utils merge conflict

* solve merge conflicts in modeling_utils.py

* add run_generation improvements from PR #2749

* adapted language generation to not use hardcoded -1 if no padding token is available

* remove the -1 removal as hard coded -1`s are not necessary anymore

* add lightweight language generation testing for randomely initialized models - just checking whether no errors are thrown

* add slow language generation tests for pretrained models using hardcoded output with pytorch seed

* delete ipdb

* check that all generated tokens are valid

* renaming

* renaming Generation -> Generate

* make style

* updated so that generate_beam_search has same token behavior than generate_no_beam_search

* consistent return format for run_generation.py

* deleted pretrain lm generate tests -> will be added in another PR

* cleaning of unused if statements and renaming

* run_generate will always return an iterable

* make style

* consistent renaming

* improve naming, make sure generate function always returns the same tensor, add docstring

* add slow tests for all lmhead models

* make style and improve example comments modeling_utils

* better naming and refactoring in modeling_utils

* improving generation

* finalized special token behaviour for no_beam_search generation

* solved modeling_utils merge conflict

* solve merge conflicts in modeling_utils.py

* add run_generation improvements from PR #2749

* adapted language generation to not use hardcoded -1 if no padding token is available

* remove the -1 removal as hard coded -1`s are not necessary anymore

* add lightweight language generation testing for randomely initialized models - just checking whether no errors are thrown

* add slow language generation tests for pretrained models using hardcoded output with pytorch seed

* delete ipdb

* check that all generated tokens are valid

* renaming

* renaming Generation -> Generate

* make style

* updated so that generate_beam_search has same token behavior than generate_no_beam_search

* consistent return format for run_generation.py

* deleted pretrain lm generate tests -> will be added in another PR

* cleaning of unused if statements and renaming

* run_generate will always return an iterable

* make style

* consistent renaming

* improve naming, make sure generate function always returns the same tensor, add docstring

* add slow tests for all lmhead models

* make style and improve example comments modeling_utils

* better naming and refactoring in modeling_utils

* changed fast random lm generation testing design to more general one

* delete in old testing design in gpt2

* correct old variable name

* temporary fix for encoder_decoder lm generation tests - has to be updated when t5 is fixed

* adapted all fast random generate tests to new design

* better warning description in modeling_utils

* better comment

* better comment and error message
Co-authored-by: Thomas Wolf <thomwolf@users.noreply.github.com>

* Added CamembertForQuestionAnswering

* fixed camembert tokenizer case

Tensorflow does not use .eval() vs .train().

closes https://github.com/huggingface/transformers/issues/2906

* Results same as fairseq
* Wrote a ton of tests
* Struggled with api signatures
* added some docs

* Removed unused fields in DistilBert TransformerBlock