Commit Graph

8557 Commits

Author SHA1 Message Date
Patrick von Platen
05fa1a7ac1 Release: v4.15.0 2021-12-22 18:43:15 +01:00
Sylvain Gugger
87a033d9fa
Properly indent return block (#14887) 2021-12-22 12:28:45 -05:00
Michael Benayoun
13504dcbea
Onnx enable tasks for supported models (part 2) (#14700)
* Revert "Revert "Added support for other features for already supported models (#14358)" (#14679)"

This reverts commit 0f4e39c559.

* is_torch_available test to avoid failing imports

* sorting parameterize parameters to solve ERROR gw0 gw1

* tests fix

* tests fix

* GPT2 with past fix

* Fixed stateful class attribute change that was breaking things when converting multiple models sequentially

* Removed onnx file

* Implemented suggestions

* Fixed __init__ to resolve conflict with master

* Remove commented import
2021-12-22 14:43:11 +01:00
Mario Šaško
1045a36c1f
Fix pytorch image classification example (#14883)
* Update example

* Remove skip in tests
2021-12-22 14:42:19 +01:00
NielsRogge
7df4b90c76
Fix Perceiver docs (#14879) 2021-12-22 14:18:03 +01:00
Sylvain Gugger
e37bc579fc Fix typo in error message 2021-12-22 08:19:36 -05:00
charon____
17efc806b4
IterableDatasetShard should use per device batch size instead of real batch size (#14714) 2021-12-22 07:52:07 -05:00
guillaume-be
2a56edb321
Updated deberta attention (#14625)
* Removed unused p2p attention handling

* Updated DeBERTa configuration

* Updated TF DeBERTa attention

* Rolled back accidental comment deletion

Co-authored-by: Lysandre Debut <lysandre@huggingface.co>
2021-12-22 07:36:08 -05:00
Ryokan RI
824fd44fc3
Feature/fix slow test in mluke (#14749)
* make MLukeTokenizerTest fast

* make LukeTokenizerTest fast

* add entry to _toctree.yaml
2021-12-22 06:35:59 -05:00
SaulLu
c94c1b8967
update the arguments add_prefix_space and trim_offsets in backend_tokenizer.post_processor of RobertaTokenizerFast (#14752)
* add tests

* change post-processor, pre-tokenizer and decoder (can't update decoder)

* update test (remove decoder which doesn't depend on trim and add_prefix)

* just update the post_processor

* fix change

* `trim_offsets` has no influence on `pre_tokenizer`

* remove a test that need some input from the `tokenizers` lib maintainers

* format

* add new test offsets roberta

* polish comments
2021-12-22 10:51:55 +01:00
Lysandre Debut
ec3567fe20
Convert model files from rst to mdx (#14865)
* First pass

* Apply suggestions from code review

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
2021-12-22 03:27:30 -05:00
Sylvain Gugger
d0422de563
Fix doc mistakes (#14874)
* Remove double returns

* Last fixes

* Quality

* Last fix for Lxmert
2021-12-21 18:54:41 -05:00
Sylvain Gugger
e846a56ca4
Fix FlaxMarianMTModel return block. (#14873)
* Fixes in marian doc

* Another time

* Add return block in FlaxMarianMTModel
2021-12-21 17:57:37 -05:00
Sylvain Gugger
a6b7b47a39
Fixes in marian doc (#14872)
* Fixes in marian doc

* Another time
2021-12-21 17:17:02 -05:00
Mishig Davaadorj
eec9c8bbd7
Fix FLAX_MULTIPLE_CHOICE_SAMPLE typo (#14871) 2021-12-21 16:54:10 -05:00
Sylvain Gugger
e51c7b5872 Skip failing test 2021-12-21 15:15:17 -05:00
Sylvain Gugger
27b3031de2
Mass conversion of documentation from rst to Markdown (#14866)
* Convert docstrings of all configurations and tokenizers

* Processors and fixes

* Last modeling files and fixes to models

* Pipeline modules

* Utils files

* Data submodule

* All the other files

* Style

* Missing examples

* Style again

* Fix copies

* Say bye bye to rst docstrings forever
2021-12-21 15:06:33 -05:00
Stas Bekman
185876392c
[doc porting] several docs (#14858)
* [doc porting] 2 docs

* [doc porting] 2 docs

* Apply suggestions from code review

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

* Update docs/source/main_classes/deepspeed.mdx

* cleanup

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
2021-12-21 09:55:25 -08:00
Stas Bekman
033c3ed95a
[examples/summarization] deal with None in data records (#14816)
* [examples/summarization] deal with None in data records

* rewrite to use a simpler (slower) variant
2021-12-21 09:17:28 -08:00
Sylvain Gugger
c075fb7855
Replace commit sha by commit url for update jobs (#14852)
* Replace commit sha by commit url for update jobs

* Typo

* Update .github/workflows/build_documentation.yml

Co-authored-by: Julien Chaumond <julien@huggingface.co>

* Apply review comments

Co-authored-by: Julien Chaumond <julien@huggingface.co>
2021-12-21 11:17:11 -05:00
Leandro von Werra
5722d05831
Add custom stopping_criteria and logits_processor to generate (#14779)
* add custom `stopping_criteria` and `logits_processor` to `generate`

* add tests for custom `stopping_criteria` and `logits_processor`

* fix typo in RAG

* address reviewer comments

* improve custom logits processor/stopping criteria error message

* fix types in merge function signature

* change default for custom list from `None` to empty list

* fix rag generate

* add string split suggestion

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
2021-12-21 16:47:41 +01:00
Zed
0062058399
Fix the value error typo of AdamW's betas' valid values checking (#14780)
* Fix the value error typo of AdamW's betas value check

* error fixed
2021-12-21 09:44:09 -05:00
Patrick von Platen
7ae6f07004
[ASR example] Improve example + add more examples (#14848)
* up

* load up

* up
2021-12-21 13:12:22 +01:00
Sylvain Gugger
97ec17f73b
Only create the model card on process 0 (#14857) 2021-12-21 06:34:47 -05:00
Patrick von Platen
b513ec8bbd
[Bart] better error message (#14854) 2021-12-21 11:57:42 +01:00
Sylvain Gugger
7af80f6618
Convert docstrings of modeling files (#14850)
* Convert file_utils docstrings to Markdown

* Test on BERT

* Return block indent

* Temporarily disable doc styler

* Remove from quality checks as well

* Remove doc styler mess

* Remove check from circleCI

* Fix typo

* Convert file_utils docstrings to Markdown

* Test on BERT

* Return block indent

* Temporarily disable doc styler

* Remove from quality checks as well

* Remove doc styler mess

* Remove check from circleCI

* Fix typo

* Let's go on all other model files

* Add templates too

* Styling and quality
2021-12-21 05:37:32 -05:00
Sylvain Gugger
2a33734606
Make the onnx submodule init lazy (#14855)
* Use lazy init for onnx submodule

* Remove debug statements
2021-12-21 03:11:25 -05:00
Stas Bekman
b6ec956976
[logging] implement warning_advice / TRANSFORMERS_NO_ADVISORY_WARNINGS (#14669)
* [logging] implement warning_advice / TRANSFORMERS_NO_ADVISORY_WARNINGS

* reword
2021-12-20 20:48:38 -08:00
Stas Bekman
c1125dc2ba
[doc] typo (#14849)
fix small typo
2021-12-20 12:20:21 -05:00
Sylvain Gugger
33f36c869f
Add a main_input_name attribute to all models (#14803)
* Add a main_input_name attribute to all models

* Fix tests

* Wtf Vs Code?

* Update src/transformers/models/imagegpt/modeling_imagegpt.py

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

* Style

* Fix copies

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
2021-12-20 11:19:08 -05:00
Henrik Holm
0940e9b242
Add 'with torch.no_grad()' to integration test forward pass (#14820) 2021-12-20 09:28:17 -05:00
Henrik Holm
b37cf7dee4
Add 'with torch.no_grad()' to integration test forward pass (#14821) 2021-12-20 09:25:34 -05:00
Patrick von Platen
952a77b05d
[Perceiver] Skip multi-gpu tests for now (#14813)
* [Perceiver] Skip multi-gpu tests for now

* Update tests/test_modeling_perceiver.py

* up

* up
2021-12-20 15:22:50 +01:00
Derek Chia
8a818c26cb
Fix dead link to benchmarks.ipynb (#14842)
Notebook has been updated here https://github.com/huggingface/notebooks/tree/master/examples/benchmark.ipynb
2021-12-20 09:08:05 -05:00
Kamal Raj
1b0ca7d270
Update CONTRIBUTING.md (#14835)
fix cmd typo
2021-12-20 08:42:03 -05:00
Chang Lan
1531b31978
Add an argument to set bucket_cap_mb for PyTorch DDP (#14756)
* [trainer] Set bucket_cap_mb for DDP from arguments

* Put find_unused_parameters into kwargs
2021-12-20 08:41:40 -05:00
Anton Lozhkov
3883e3a75e
Add SD and SV heads for WavLM (#14847)
* Add converted heads

* Add dummies
2021-12-20 16:40:56 +03:00
Patrick von Platen
cd583bdaa5
[WavLM] Fix slow tests (#14845) 2021-12-20 12:06:42 +01:00
Patrick von Platen
281e1fba75
up (#14829) 2021-12-20 11:47:32 +01:00
Patrick von Platen
091693b494
[Seq2SeqTrainer] Remove model input name hack (#14802)
* [Seq2SeqTrainer] Remove model input name hack

* Update src/transformers/trainer_seq2seq.py

* make style

* finish
2021-12-20 10:53:48 +01:00
Patrick von Platen
84ea427f46
[ImageGPT] Deprecate pixel_values input name to input_ids (#14801)
* [ImageGPT] Deprecate pixel_values input name to input_ids

* up

* Apply suggestions from code review

Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>

* correct

* finish

Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>
2021-12-17 20:05:22 +01:00
Patrick von Platen
c4a96cecbc
Wav2Vec2 meets phonemes (#14353)
* up

* add tokenizer

* improve more

* finish tokenizer

* finish

* adapt speech recognition script

* adapt convert

* more fixes

* more fixes

* update phonemizer wav2vec2

* better naming

* fix more tests

* more fixes swedish

* correct tests

* finish

* improve script

* remove file

* up

* lets get those 100 model architectures until the end of the month

* make fix-copies

* correct more

* correct script

* more fixes

* more fixes

* add to docs

* Apply suggestions from code review

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

* replace assert

* fix copies

* fix docs

* new try docs

* boom boom

* update

* add phonemizer to audio tests

* make fix-copies

* up

* upload models

* some changes

* Update tests/test_tokenization_wav2vec2_phoneme.py

Co-authored-by: Anton Lozhkov <aglozhkov@gmail.com>

* more fixes

* remove @

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: Anton Lozhkov <aglozhkov@gmail.com>
2021-12-17 19:56:44 +01:00
Lysandre Debut
77d6c826d8
Convert rst to mdx bert (#14806)
* BERT to mdx
mdx :)
c

* Update docs/source/model_doc/bert.mdx

Co-authored-by: Julien Chaumond <julien@huggingface.co>

* Remove all
Co-authored-by: sgugger <sylvain.gugger@gmail.com>

Co-authored-by: Julien Chaumond <julien@huggingface.co>
2021-12-17 11:13:34 -05:00
Sylvain Gugger
0b4ea79a0c Trigger doc building 2021-12-17 11:14:18 -05:00
Daniel Stancl
ff066119ca
Implement head_mask for Flax BERT and other models copied from BERT (#14620)
* Implement head_mask for Flax BERT and other models copied from BERT

* Remove `from jax._src.nn.functions import sigmoid`

Remove `from jax._src.nn.functions import sigmoid` unintentionally added by IDE

* Remove no more valid copy statement

* Apply patil-suraj's suggestions from code review

* Apply suggestions from the code review

* Update Flax template

* Fix a typo

* Also update template for CausalLM modules
2021-12-17 17:06:59 +01:00
Patrick von Platen
95119ad7b0
[Generate] Correct input_ids detection (#14815)
* [Generate] Correct input_ids detection

* correct
2021-12-17 16:08:54 +01:00
Patrick von Platen
bdbe3df869
[WavLM] Layerdrop is not allowed for first layer (#14811)
* [WavLM] Layerdrop is not allowed for first layer

* Apply suggestions from code review
2021-12-17 13:30:18 +01:00
NielsRogge
cbf036f7ae
Add test (#14810) 2021-12-17 04:33:27 -05:00
Patrick von Platen
c4a0fb5199
[WavLM] Correct position bias computation (#14805) 2021-12-16 22:42:57 +01:00
Lysandre Debut
d194d639ab
Remove datasets requirement (#14795) 2021-12-16 14:34:14 -05:00