Commits · 956ae62139cf454d6a62daaf1e332aa2e06ec15f · chenpangpang / transformers

"tools/git@developer.sourcefind.cn:gaoqiong/migraphx.git" did not exist on "15a7d96a10ae9d3a8856347aeaa23a47a1ba61ef"

03 Mar, 2023 12 commits
- Fix wrong documentation about DataCollator padding defaults (#21919) · 956ae621
  substanc3 authored Mar 03, 2023
```
* Fix wrong documentation about DataCollator padding defaults

* Fix styling
```
  956ae621
- Avoid failure in `check_repo.py` due to missing backends (#21930) · 8c40ba73
  Yih-Dar authored Mar 03, 2023
```
* Update utils/check_repo.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update utils/check_repo.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

---------
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
```
  8c40ba73
- Fix `AlignModelTest` tests (#21923) · d4306dae
  Yih-Dar authored Mar 03, 2023
```
* fix

* fix

---------
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
```
  d4306dae
- feat: filter try/except when looking at custom code (#21914) · c5a1ff9e
  Zach Nussbaum authored Mar 03, 2023
```
* feat: filter try/except

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

---------
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
```
  c5a1ff9e
- Cleanup more auto mapping names (#21909) · 02a77fa0
  Yih-Dar authored Mar 03, 2023
```
* fix auto 2

* fix auto 2

* fix task guide issue

* fix

---------
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
```
  02a77fa0
- Use large VM for `repo_utils_job` (#21928) · b05e0bec
  Yih-Dar authored Mar 03, 2023
```
upgrade to large VM
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
```
  b05e0bec
- Update `model_split_percents` for `WhisperModelTest` (#21922) · fa9d2ad7
  Yih-Dar authored Mar 03, 2023
```
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
```
  fa9d2ad7
- Fix gradient checkpointing megatron bert (#21921) · c82bd371
  Karim Foda authored Mar 03, 2023
  
  c82bd371
- Fix gradient checkpointing bug in mvp (#21920) · 99a62347
  Karim Foda authored Mar 03, 2023
  
  99a62347
- Fix gradient checkpointing bug in MBart (#21918) · e407b5a3
  Karim Foda authored Mar 03, 2023
  
  e407b5a3
- faster forward following what is done for images (#21906) · dcec3277
  Arthur authored Mar 03, 2023
```
* faster forward following what is done for images

* add missing licence
```
  dcec3277
- Fix doctests for TFVisionTextDualEncoder (#21910) · 37e0974a
  Matt authored Mar 03, 2023
  
  37e0974a
02 Mar, 2023 23 commits

Avoid modeling tests run in pipeline CI jobs (#21911) · 9f5bfe1b

Yih-Dar authored Mar 02, 2023



* rework is_pipeline_test

* bring back 3 tests

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

9f5bfe1b

[time series] Add Time series inputs tests (#21846) · db979f75

Kashif Rasul authored Mar 02, 2023



* intial test of inputs

* added test for generation

* remove asserts

* fixed test

* Update tests/models/time_series_transformer/test_modeling_time_series_transformer.py
Co-authored-by: NielsRogge <48327001+NielsRogge@users.noreply.github.com>

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

db979f75

Faster zero shot image (#21897) · b2a41d2b

Nicolas Patry authored Mar 02, 2023



* Make ZeroShotImageClassificationPipeline faster

The pipeline makes separate calls to model for each candidate label.
This commit combines all labels into one call.
Original code takes more that 60 seconds to process one image and 1000
candidate labels. Updated code takes less than 2 seconds.

* implement batching

* code formatting

* Creating an even faster zero-shot-image-classifiction.

Unfortunately super tailored towards CLIP.
Co-Authored-By: Yessen Kanapin <yessen@deepinfra.com>

* Quality.

* Cleanup.

* Order different on the CI it seems.

* Cleanup.

* Quality.

---------
Co-authored-by: Yessen Kanapin <yessen@deepinfra.com>

b2a41d2b

Temporarily skip 3 tests in `BridgeTowerModelTest` (#21908) · 88e5c51a
Yih-Dar authored Mar 02, 2023
```
skip for now
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
```
88e5c51a

Add Blip and Blip2 for pipeline tests (#21904) · e6de9186

Yih-Dar authored Mar 02, 2023



* fix

* add to tests

* style and quality

* add missing

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

e6de9186

Refactor whisper asr pipeline to include language too. (#21427) · 13254591

Nicolas Patry authored Mar 02, 2023



* [WIP] whisper refacto to support language output.

* Handling merges.

* A bit more cleanup and comments.

* Many improvements.

Lots of details everywhere.

* Cleanup old code and tests.

* Handle lone timestamp tokens (just recover when something bad happens).

* Adding return_language example.

* No ffmpeg.

* Hmm.

* Some corrections.

* Both fast and slow.

* New black.

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

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

* Remove print.

* Undoing tests modifications.

* Smaller test modifications.

* Rename.

* Remove maxDiff.

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

13254591

Make schedulers picklable by making lr_lambda fns global (#21768) · 8e5a1b2a

Connor Henderson authored Mar 02, 2023

* Make schedulers picklable by making lr_lambda fns global

* add unused _get_constant_schedule_lr_lambda arg

* remove unneeded _get_constant_schedule_lr_lamda

* add test

* make style

* rebase, remove torch dep, put lambda back

* repo-consistency and style

8e5a1b2a

Prophetnet batch dimension inversion fix (#21870) · 6bf88537

Kian Sierra McGettigan authored Mar 02, 2023

* decoder forward pass is working

* no model has forward pass returning attentions

* decoder ngram changed to not mix batch size

* current basic forward pass returns identical result

* passed test_model attentions

* passed test_encoder_decoder_model_generate

* passed test_headmasking

* removed old block

* removed comments bug/fixme

* removed bug comments

* applied styling

* applied fix-copies

* applied ngram forward comments

* corrected dimension notation

* applied styling and comment fixes

* changed asserts for raise ValueError

* changed question gen test

* updated hidden_states integration test

* applied styling

6bf88537

Clean up auto mapping names (#21903) · 99ba36e7

Yih-Dar authored Mar 02, 2023



* add new test

* fix after new test

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

99ba36e7

Mark pipeline tests to skip them easily (#21887) · 50a8ed3e

Sylvain Gugger authored Mar 02, 2023



* Mark pipeline tests to skip them easily

* Mark the mixin as pipeline test

* Update src/transformers/testing_utils.py
Co-authored-by: Yih-Dar <2521628+ydshieh@users.noreply.github.com>

---------
Co-authored-by: Yih-Dar <2521628+ydshieh@users.noreply.github.com>

50a8ed3e

Fix gradient checkpointing bug marian (#21842) · d9e28d91
Karim Foda authored Mar 02, 2023
```
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
```
d9e28d91
Fix gradient checkpointing bug M2M 100 (#21841) · b405b62f
Karim Foda authored Mar 02, 2023
```
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
```
b405b62f
Fix gradient checkpointing bug LED (#21840) · 7e6dd664
Karim Foda authored Mar 02, 2023
```
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
```
7e6dd664
fsdp bf16 enable autocast (#21847) · b6f47b53
Sourab Mangrulkar authored Mar 02, 2023

b6f47b53
[GPT-J] add deprecation warning (#21869) · fb76994c
Arthur authored Mar 02, 2023
```
* add deprecation warning

* remove pos ids from args docstirng

* fix failing test
```
fb76994c
fix typo in Bart's attention (#21898) · 648d0deb
Kashif Rasul authored Mar 02, 2023

648d0deb

[Whisper] Add rescaling function with `do_normalize` (#21263) · c87654dc

Arthur authored Mar 02, 2023



* add `zero_mean_unit_var_norm` function

* normalize before MEL computation

* fixup

* add simple test

* quality

* Update tests/models/whisper/test_feature_extraction_whisper.py
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>

* fixup

* use attention masks if padding was applied

* Update based on review
Co-authored-by: bofeng huang <bofenghuang7@gmail.com>

---------
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>
Co-authored-by: bofeng huang <bofenghuang7@gmail.com>

c87654dc

[T5 doc] Fix confusing documentation about `d_kv` (#21896) · b48c7f7b
Arthur authored Mar 02, 2023
```
* Confusing documentation in T5

* Fix onfusing documentation in T5 configuration file
```
b48c7f7b
Add `inputs_embeds` functionality when generating with BioGPT (#21889) · edbb37f7
Sid Kiblawi authored Mar 02, 2023
```
* initial commit to add inputs_embeds to generation

* formatting
```
edbb37f7

Use PyAV instead of Decord in examples (#21572) · 3412f597

amyeroberts authored Mar 02, 2023

* Use PyAV instead of Decord

* Get frame indices

* Fix number of frames

* Update src/transformers/models/videomae/image_processing_videomae.py

* Fix up

* Fix copies

* Update timesformer doctests

* Update docstrings

3412f597

[ZAC] fix ci daily (#21893) · c256bc6d
Arthur authored Mar 02, 2023
```
add correct revision after model was overwritten
```
c256bc6d

[Refactor] Relative imports wherever we can (#21880) · 633e5e89

Arthur authored Mar 02, 2023

* initial commit

* update

* second batch

* style

* fix imports

* fix relative import on pipeline

633e5e89

fix checkpoint (#21874) · 43299c63
Arthur authored Mar 02, 2023

43299c63

01 Mar, 2023 5 commits

Fix `test_load_default_pipelines_pt` for `ClapModel` (#21886) · 89359e4c
Yih-Dar authored Mar 01, 2023
```
* fix tests

---------
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
```
89359e4c

Fix `WhisperModelTest` (#21883) · 36ee1283

Yih-Dar authored Mar 01, 2023



* force on the same device

* fix tests

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

36ee1283

Fix Gradient checkpointing bug BigBird (#21882) · 4edfd2d4
saswatmeher authored Mar 02, 2023
```
Co-authored-by: saswatmeher <saswatmeher@cse.iitb.ac.in>
```
4edfd2d4

Add ALIGN to transformers (#21741) · 269b0549

Alara Dirik authored Mar 01, 2023

Adds the ALIGN model to transformers. ALIGN is introduced in "Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision" by Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc V. Le, Yunhsuan Sung, Zhen Li, Tom Duerig.

269b0549

Add TFVisionTextDualEncoder (#21873) · f7c618e3

Matt authored Mar 01, 2023



* Temporary commit to stash everything so far

* Temporary commit to stash everything so far

* stash commit

* Refactor from_pretrained

* Fix final test, make fixup

* Update dummies

* Add model to TEST_FILES_WITH_NO_COMMON_TESTS

* Update src/transformers/models/vision_text_dual_encoder/modeling_tf_vision_text_dual_encoder.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Update src/transformers/models/vision_text_dual_encoder/modeling_tf_vision_text_dual_encoder.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Update src/transformers/models/vision_text_dual_encoder/modeling_tf_vision_text_dual_encoder.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Update src/transformers/models/vision_text_dual_encoder/modeling_tf_vision_text_dual_encoder.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Add TFVisionTextDualEncoder to utils/documentation_tests.txt

* make fixup

---------
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

f7c618e3