0c55a384f8
Add reference to NLP dataset ( #5028 )
...
* Add reference to NLP dataset
* Update README.md
Co-authored-by: Julien Chaumond <chaumond@gmail.com>
2020-06-16 04:19:09 -04:00
0946d1209d
Add reference to NLP (package) dataset ( #5029 )
...
* Add reference to NLP (package) dataset
* Update README.md
Co-authored-by: Julien Chaumond <chaumond@gmail.com>
2020-06-16 04:17:46 -04:00
edcb3ac59a
refactor(wandb): consolidate import ( #5044 )
2020-06-16 03:40:43 -04:00
9e03364999
Ability to pickle/unpickle BatchEncoding pickle (reimport) ( #5039 )
...
* Added is_fast property on BatchEncoding to indicate if the object comes from a Fast Tokenizer.
* Added __get_state__() & __set_state__() to be pickable.
* Correct tokens() return type from List[int] to List[str]
* Added unittest for BatchEncoding pickle/unpickle
* Added unittest for BatchEncoding is_fast
* More careful checking on BatchEncoding unpickle tests.
* Formatting.
* is_fast should assertTrue on Rust tokenizers.
* Ensure tensorflow has correct way of checking array_equal
* More formatting.
2020-06-16 09:25:25 +02:00
f9f8a5312e
Add DistilBertForMultipleChoice ( #5032 )
...
* Add `DistilBertForMultipleChoice`
2020-06-15 18:31:41 -04:00
36434220fc
[HUGE] Refactoring tokenizers backend - padding - truncation - pre-tokenized pipeline - fast tokenizers - tests ( #4510 )
...
* Use tokenizers pre-tokenized pipeline
* failing pretrokenized test
* Fix is_pretokenized in python
* add pretokenized tests
* style and quality
* better tests for batched pretokenized inputs
* tokenizers clean up - new padding_strategy - split the files
* [HUGE] refactoring tokenizers - padding - truncation - tests
* style and quality
* bump up requied tokenizers version to 0.8.0-rc1
* switched padding/truncation API - simpler better backward compat
* updating tests for custom tokenizers
* style and quality - tests on pad
* fix QA pipeline
* fix backward compatibility for max_length only
* style and quality
* Various cleans up - add verbose
* fix tests
* update docstrings
* Fix tests
* Docs reformatted
* __call__ method documented
Co-authored-by: Thomas Wolf <thomwolf@users.noreply.github.com>
Co-authored-by: Lysandre <lysandre.debut@reseau.eseo.fr>
2020-06-15 17:12:51 -04:00
ebba39e4e1
[Bart] Question Answering Model is added to tests ( #5024 )
...
* fix test
* Update tests/test_modeling_common.py
* Update tests/test_modeling_common.py
2020-06-15 22:50:09 +02:00
bbad4c6989
Add position_ids ( #5021 )
2020-06-15 15:50:17 -04:00
1bf4098e03
feat(TFTrainer): improve logging ( #4946 )
...
* feat(tftrainer): improve logging
* fix(trainer): consider case with evaluation only
* refactor(tftrainer): address comments
* refactor(tftrainer): move self.epoch_logging to __init__
2020-06-15 14:06:17 -04:00
7b5a1e7d51
Fix importing transformers on Windows ( #4997 )
2020-06-15 19:36:57 +02:00
a9f1fc6c94
Add bart-base ( #5014 )
2020-06-15 13:29:26 -04:00
7b685f5229
Increase pipeline support for ONNX export. ( #5005 )
...
* Increase pipeline support for ONNX export.
* Style.
2020-06-15 19:13:58 +02:00
1affde2f10
Make DataCollator a callable ( #5015 )
...
* Make DataCollator a callable
* Update src/transformers/data/data_collator.py
Co-authored-by: Julien Chaumond <chaumond@gmail.com>
2020-06-15 11:58:33 -04:00
f7c93b3cee
Possible fix to make AMP work with DDP in the trainer ( #4728 )
...
* manually set device in trainer args
* check if current device is cuda before set_device
* Explicitly set GPU ID when using single GPU
This addresses https://github.com/huggingface/transformers/issues/4657#issuecomment-642228099
2020-06-15 10:10:26 -04:00
66bcfbb130
Create README.md ( #4975 )
...
* Create README.md
* Update model_cards/ipuneetrathore/bert-base-cased-finetuned-finBERT/README.md
Co-authored-by: Julien Chaumond <chaumond@gmail.com>
2020-06-15 08:43:50 -04:00
d812e6d76e
NER: fix construction of input examples for RoBERTa ( #4943 )
...
* utils_ner: do not add extra sep token for RoBERTa model
* run_pl_ner: do not add extra sep token for RoBERTa model
2020-06-15 08:30:40 -04:00
ebab096e86
[model card] model card for bart-large-finetuned-squadv1 ( #4977 )
...
* [model card] model card for bart-large-finetuned-squadv1
* add metadata link to the dataset
2020-06-15 05:39:41 -04:00
9ad36ad57f
Improve ONNX logging ( #4999 )
...
* Improve ONNX export logging to give more information about the generated graph.
* Correctly handle input and output in the logging.
2020-06-15 11:04:51 +02:00
9931f817b7
fix ( #4976 )
2020-06-14 21:36:14 +02:00
9208f57b16
BartTokenizerFast ( #4878 )
2020-06-14 13:04:49 -04:00
403d309857
Hans data ( #4854 )
...
* Update hans data to be able to use Trainer
* Fixes
* Deal with tokenizer that don't have token_ids
* Clean up things
* Simplify data use
* Fix the input dict
* Formatting + proper path in README
2020-06-13 09:35:13 -04:00
ca5e1cdf8e
model_cards: we can now tag datasets
...
see corresponding model pages to see how it's rendered
2020-06-12 23:19:07 +02:00
e93ccb3290
BartForQuestionAnswering ( #4908 )
2020-06-12 15:47:57 -04:00
538531cde5
Add AlbertForMultipleChoice ( #4959 )
...
* Add AlbertForMultipleChoice
* Make up to date and add all models to common tests
2020-06-12 14:20:19 -04:00
fe24139702
Create README.md ( #4865 )
2020-06-12 09:03:43 -04:00
9aa219a1fe
Create README.md ( #4872 )
2020-06-12 09:03:13 -04:00
86578bb04c
[AutoModel] Split AutoModelWithLMHead into clm, mlm, encoder-decoder ( #4933 )
...
* first commit
* add new auto models
* better naming
* fix bert automodel
* fix automodel for pretraining
* add models to init
* fix name typo
* fix typo
* better naming
* future warning instead of depreciation warning
2020-06-12 10:01:49 +02:00
5620033115
[mbart] Fix fp16 testing logic ( #4949 )
2020-06-11 22:11:34 -04:00
473808da0d
update `mvmt-pruning/saving_prunebert` (updating torch to 1.5)
2020-06-11 19:42:45 +00:00
caf3746678
fix indentation issue ( #4941 )
2020-06-11 21:28:01 +02:00
6293eb04df
[Model card] model card for electra-base QA model ( #4936 )
2020-06-11 13:16:34 -04:00
08b59d10e5
MBartTokenizer:add language codes ( #3776 )
2020-06-11 13:02:33 -04:00
20451195f0
Support multiple choice in tf common model tests ( #4920 )
...
* Support multiple choice in tf common model tests
* Add the input_embeds test
2020-06-11 10:31:26 -04:00
699541c4b3
TFTrainer: Add dataloader_drop_last ( #4925 )
2020-06-11 02:11:22 -04:00
e80d6c689b
Fix resize_token_embeddings for Transformer-XL ( #4759 )
...
* Fixed resize_token_embeddings for transfo_xl model
* Fixed resize_token_embeddings for transfo_xl.
Added custom methods to TransfoXLPreTrainedModel for resizing layers of
the AdaptiveEmbedding.
* Updated docstring
* Fixed resizinhg cutoffs; added check for new size of embedding layer.
* Added test for resize_token_embeddings
* Fixed code quality
* Fixed unchanged cutoffs in model.config
Co-authored-by: Rafael Weingartner <rweingartner.its-b2015@fh-salzburg.ac.at>
2020-06-10 19:03:06 -04:00
d541938c48
Make multiple choice models work with input_embeds ( #4921 )
2020-06-10 18:38:34 -04:00
1e2631d6f8
Split LMBert model in two ( #4874 )
...
* Split LMBert model in two
* Fix example
* Remove lm_labels
* Adapt tests, refactor prepare_for_generation
* Fix merge
* Hide BeartLMHeadModel
2020-06-10 18:26:42 -04:00
f6da8b2200
check type before logging in trainer to ensure values are scalars ( #4883 )
...
* check type before logging to ensure it's a scalar
* log when Trainer attempts to add a non-scalar value using TensorboardX's writer.add_scalar so we know what kinds of fixes are appropriate
* black it
* rephrase log message to clarify attribute was dropped
Co-authored-by: Julien Chaumond <chaumond@gmail.com>
Co-authored-by: Julien Chaumond <chaumond@gmail.com>
2020-06-10 18:25:55 -04:00
1c986f42ff
Create README.md ( #4871 )
2020-06-10 17:29:41 -04:00
3ae2e86baf
Run a single wandb instance per TPU run ( #4851 )
...
* Run a single wandb instance per TPU run
* wandb: self.is_world_master
* make style
Co-authored-by: Julien Chaumond <chaumond@gmail.com>
2020-06-10 16:28:18 -04:00
466aa57a45
Don't init TPU device twice ( #4916 )
2020-06-10 15:53:15 -04:00
ef2dcdccaa
ElectraForQuestionAnswering ( #4913 )
...
* ElectraForQuestionAnswering
* udate __init__
* add test for electra qa model
* add ElectraForQuestionAnswering in auto models
* add ElectraForQuestionAnswering in all_model_classes
* fix outputs, input_ids defaults to None
* add ElectraForQuestionAnswering in docs
* remove commented line
2020-06-10 15:17:52 -04:00
5d63ca6c38
[ctrl] fix pruning of MultiHeadAttention ( #4904 )
2020-06-10 14:06:55 -04:00
4e10acb3e5
Add more models to common tests ( #4910 )
2020-06-10 13:19:53 -04:00
3b3619a327
[All models] fix docs after adding output attentions to all forward functions ( #4909 )
...
* fix doc
* add format file
* add output attentions to all docs
* add also for bart
* fix naming
* re-add doc to config
2020-06-10 18:10:59 +02:00
ac99217e92
Fix the CI ( #4903 )
...
* Fix CI
2020-06-10 09:26:06 -04:00
0a375f5abd
Deal with multiple choice in common tests ( #4886 )
...
* Deal with multiple choice in common tests
2020-06-10 08:10:20 -04:00
e8db8b845a
Remove unused arguments in Multiple Choice example ( #4853 )
...
* Remove unused arguments
* Formatting
* Remove second todo comment
2020-06-09 20:05:09 -04:00
29c36e9f36
run_pplm.py bug fix ( #4867 )
...
`is_leaf` may become `False` after `.to(device=device)` function call.
2020-06-09 19:14:27 -04:00
13aa174112
uninstalled wandb raises AttributeError
2020-06-09 18:50:56 -04:00
Manuel Romero
Boris Dayma
Funtowicz Morgan
Sylvain Gugger
Anthony MOI
Patrick von Platen
Sam Shleifer
Bram Vanroy
ipuneetrathore
Stefan Schweter
Suraj Patil
ZhuBaohe
Julien Chaumond
Yannis Papanikolaou
VictorSanh
Setu Shah
RafaelWO
Matthew Goldey
Lysandre Debut
Amil Khare
songyouwei
Lysandre