Skip to content

GitLab

Switch branch/tag
  1. 10 Aug, 2022 1 commit
  3. 09 Aug, 2022 1 commit
  5. 08 Aug, 2022 2 commits
    • Sylvain Gugger's avatar
      Clean up hub (#18497) · 377cdded
      Sylvain Gugger authored 
      * Clean up utils.hub

* Remove imports

* More fixes

* Last fix
      377cdded
    • Nicolas Patry's avatar
      [DX fix] Fixing QA pipeline streaming a dataset. (#18516) · a4562552
      Nicolas Patry authored 
      * [DX fix] Fixing QA pipeline streaming a dataset.

QuestionAnsweringArgumentHandler would iterate over the whole dataset
effectively killing all properties of the pipeline.
This restores nice properties when using `Dataset` or `Generator` since
those are meant to be consumed lazily.

* Handling TF better.
      a4562552
  7. 06 Aug, 2022 1 commit
  9. 05 Aug, 2022 2 commits
    • Sylvain Gugger's avatar
      Fix pipeline tests (#18487) · 70fa1a8d
      Sylvain Gugger authored 
      * Fix pipeline tests

* Make sure all pipelines tests run with init changes
      70fa1a8d
    • Nicolas Patry's avatar
      Fixing issue where generic model types wouldn't load properly with the pipeline (#18392) · 586dcf6b
      Nicolas Patry authored 
      * Adding a better error message when the model is improperly configured

within transformers.

* Update src/transformers/pipelines/__init__.py

* Black version.

* Overriding task aliases so that tokenizer+feature_extractor

values are correct.

* Fixing task aliases by overriding their names early

* X.

* Fixing feature-extraction.

* black again.

* Normalizing `translation` too.

* Fixing last few corner cases.

translation need to use its non normalized name (translation_XX_to_YY,
so that the task_specific_params are correctly overloaded).
This can be removed and cleaned up in a later PR.

`speech-encode-decoder` actually REQUIRES to pass a `tokenizer` manually
so the error needs to be discarded when the `tokenizer` is already
there.

* doc-builder fix.

* Fixing the real issue.

* Removing dead code.

* Do not import the actual config classes.
      586dcf6b
  11. 03 Aug, 2022 1 commit
    • LSinev's avatar
      Fix torch version comparisons (#18460) · 02b176c4
      LSinev authored 
      Comparisons like
version.parse(torch.__version__) > version.parse("1.6")
are True for torch==1.6.0+cu101 or torch==1.6.0+cpu

version.parse(version.parse(torch.__version__).base_version) are preferred (and available in pytorch_utils.py
      02b176c4
  13. 02 Aug, 2022 2 commits
  15. 28 Jul, 2022 1 commit
  17. 26 Jul, 2022 1 commit
  19. 19 Jul, 2022 1 commit
  21. 15 Jul, 2022 2 commits
  23. 30 Jun, 2022 2 commits
  25. 28 Jun, 2022 1 commit
    • Nicolas Patry's avatar
      Fixing a regression with `return_all_scores` introduced in #17606 (#17906) · 776855c7
      Nicolas Patry authored 
      Fixing a regression with `return_all_scores` introduced in #17606

- The legacy test actually tested `return_all_scores=False` (the actual
  default) instead of `return_all_scores=True` (the actual weird case).

This commit adds the correct legacy test and fixes it.

Tmp legacy tests.

Actually fix the regression (also contains lists)

Less diffed code.
      776855c7
  27. 13 Jun, 2022 1 commit
  29. 10 Jun, 2022 1 commit
    • Balaji's avatar
      Fixes #17128 . (#17356) · 35b16032
      Balaji authored 
      

VisibleDeprecationWarning is addressed by specifying dtype=object when creating numpy array.
Update code based on review feedback.
Undo whitespace changes to tokenization_utils_base.py.
Co-authored-by: default avatarI like data <ilikedata@nym.hush.com>
      35b16032
  31. 09 Jun, 2022 2 commits
    • Nicolas Patry's avatar
      Running a pipeline of `float16`. (#17637) · c38f4e1f
      Nicolas Patry authored 
      When we're preparing the tensors for CPU for postprocessing, we need
to upgrade the `float16` to `float32` since CPUs don't have instructions
for `[b]float16`.
      c38f4e1f
    • Nicolas Patry's avatar
      Adding `top_k` argument to `text-classification` pipeline. (#17606) · 2351729f
      Nicolas Patry authored 
      * Adding `top_k` and `sort` arguments to `text-classification` pipeline.

- Deprecate `return_all_scores` as `top_k` is more uniform with other
  pipelines, and a superset of what `return_all_scores` can do.
  BC is maintained though.
  `return_all_scores=True` -> `top_k=None`
  `return_all_scores=False` -> `top_k=1`

- Using `top_k` will imply sorting the results, but using no argument
  will keep the results unsorted for backward compatibility.

* Remove `sort`.

* Fixing the test.

* Remove bad doc.
      2351729f
  33. 01 Jun, 2022 1 commit
  35. 24 May, 2022 1 commit
  37. 19 May, 2022 1 commit
  39. 18 May, 2022 2 commits
  41. 12 May, 2022 1 commit
  43. 10 May, 2022 1 commit
  45. 26 Apr, 2022 1 commit
  47. 25 Apr, 2022 2 commits
  49. 21 Apr, 2022 2 commits
  51. 20 Apr, 2022 1 commit
  53. 19 Apr, 2022 1 commit
  55. 14 Apr, 2022 1 commit
  57. 12 Apr, 2022 1 commit
    • Nicolas Patry's avatar
      Change the chunk_iter function to handle (#16730) · a192f61e
      Nicolas Patry authored 
      * Change the chunk_iter function to handle

the subtle cases where the last chunk gets ignored since all the
data is in the `left_strided` data.

We need to remove the right striding on the previous item.

* Remove commented line.
      a192f61e
  59. 25 Mar, 2022 1 commit
  61. 23 Mar, 2022 1 commit
    • Sylvain Gugger's avatar
      Reorganize file utils (#16264) · 4975002d
      Sylvain Gugger authored 
      * Split file_utils in several submodules

* Fixes

* Add back more objects

* More fixes

* Who exactly decided to import that from there?

* Second suggestion to code with code review

* Revert wront move

* Fix imports

* Adapt all imports

* Adapt all imports everywhere

* Revert this import, will fix in a separate commit
      4975002d