Skip to content

GitLab

Switch branch/tag
  1. 17 Sep, 2021 3 commits
  3. 16 Sep, 2021 4 commits
  5. 15 Sep, 2021 1 commit
  7. 14 Sep, 2021 2 commits
    • Bhadresh Savani's avatar
      [Flax] Addition of FlaxPegasus (#13420) · c1e47bf4
      Bhadresh Savani authored 
      

* added initial files

* fixes pipeline

* fixes style and quality

* fixes doc issue and positional encoding

* fixes layer norm and test

* fixes quality issue

* fixes code quality

* removed extra layer norm

* added layer norm back in encoder and decoder

* added more code copy quality checks

* update tests

* Apply suggestions from code review

* fix import

* fix test
Co-authored-by: default avatarpatil-suraj <surajp815@gmail.com>
      c1e47bf4
    • Sylvain Gugger's avatar
      Push to hub when saving checkpoints (#13503) · 3081d386
      Sylvain Gugger authored 
      * Push to hub when saving checkpoints

* Add model card

* Revert partial model card

* Small fix for checkpoint

* Add tests

* Add documentation

* Fix tests

* Bump huggingface_hub

* Fix test
      3081d386
  9. 13 Sep, 2021 3 commits
  11. 10 Sep, 2021 3 commits
    • Suraj Patil's avatar
      [GPT-Neo] Simplify local attention (#13491) · 010965dc
      Suraj Patil authored 
      * simplify local attention

* update tests

* add a comment and use torch.bitwise_xor
      010965dc
    • Patrick von Platen's avatar
      [Wav2Vec2] Fix normalization for non-padded tensors (#13512) · d7b3b709
      Patrick von Platen authored 
      * finalize

* Apply suggestions from code review

* finish cleaner implementation

* more tests

* small fix

* finish

* up
      d7b3b709
    • Nicolas Patry's avatar
      [Large PR] Entire rework of pipelines. (#13308) · c63fcabf
      Nicolas Patry authored 
      

* Enabling dataset iteration on pipelines.

Enabling dataset iteration on pipelines.

Unifying parameters under `set_parameters` function.

Small fix.

Last fixes after rebase

Remove print.

Fixing text2text `generate_kwargs`

No more `self.max_length`.

Fixing tf only conversational.

Consistency in start/stop index over TF/PT.

Speeding up drastically on TF (nasty bug where max_length would increase
a ton.)

Adding test for support for non fast tokenizers.

Fixign GPU usage on zero-shot.

Fix working on Tf.

Update src/transformers/pipelines/base.py
Co-authored-by: default avatarSylvain Gugger <35901082+sgugger@users.noreply.github.com>

Update src/transformers/pipelines/base.py
Co-authored-by: default avatarSylvain Gugger <35901082+sgugger@users.noreply.github.com>

Small cleanup.

Remove all asserts + simple format.

* Fixing audio-classification for large PR.

* Overly explicity null checking.

* Encapsulating GPU/CPU pytorch manipulation directly within `base.py`.

* Removed internal state for parameters of the  pipeline.

Instead of overriding implicitly internal state, we moved
to real named arguments on every `preprocess`, `_forward`,
`postprocess` function.

Instead `_sanitize_parameters` will be used to split all kwargs
of both __init__ and __call__ into the 3 kinds of named parameters.

* Move import warnings.

* Small fixes.

* Quality.

* Another small fix, using the CI to debug faster.

* Last fixes.

* Last fix.

* Small cleanup of tensor moving.

* is not None.

* Adding a bunch of docs + a iteration test.

* Fixing doc style.

* KeyDataset = None guard.

* RRemoving the Cuda test for pipelines (was testing).

* Even more simple iteration test.

* Correct import .

* Long day.

* Fixes in docs.

* [WIP] migrating object detection.

* Fixed the target_size bug.

* Fixup.

* Bad variable name.

* Fixing `ensure_on_device` respects original ModelOutput.
      c63fcabf
  13. 09 Sep, 2021 4 commits
  15. 08 Sep, 2021 4 commits
  17. 07 Sep, 2021 2 commits
  19. 06 Sep, 2021 5 commits
  21. 02 Sep, 2021 5 commits
    • Nathan Raw's avatar
      ✨ Add PyTorch image classification example (#13134) · 76c4d8bf
      Nathan Raw authored 
      * ✨ add pytorch image classification example

* 🔥 remove utils.py

* 💄 fix flake8 style issues

* 🔥 remove unnecessary line

* ✨ limit dataset sizes

* 📌 update reqs

* 🎨 restructure - use datasets lib

* 🎨 import transforms directly

* 📝 add comments

* 💄 style

* 🔥 remove flag

* 📌 update requirement warning

* 📝 add vision README.md

* 📝 update README.md

* 📝 update README.md

* 🎨 add image-classification tag to model card

* 🚚 rename vision ➡️ image-classification

* 📝 update image-classification README.md
      76c4d8bf
    • Patrick von Platen's avatar
      up (#13396) · 9bd5d97c
      Patrick von Platen authored
      9bd5d97c
    • Patrick von Platen's avatar
      fix (#13395) · efa4f5f0
      Patrick von Platen authored
      efa4f5f0
    • Apoorv Garg's avatar
      Correct order of overflowing_tokens for slow tokenizer (#13179) · b91e65af
      Apoorv Garg authored 
      * correct order of overflowing_tokens for slow tokenizer (issue fix #13148)

* python 3.9 requires sentencepiece version 0.1.94 or above

* slicing of ids fixed in truncated_sequence()

* Update setup.py

* Correct order of overflowing tokens for pair of sentences

* code reformatted

* Update tokenization_utils_base.py

* reformatting file

* test to check single_input added

* missing function restored

* test to check pair_input overflowing tokens order

* test to check pair_input overflowing tokens order

* test to check pair_input overflowing tokens order

* added an error message for pair of seq and longest_first strategy

* test for pair_input modified

* variable name corrected

* fixed a typo in error message

* requested changes implemented

* required test added

* Corrected the message to match test message

* added error message for Luke Tokenizer

* lost test recovered

* docstring for truncate_sequences and prepare_for_model updated

* docstring for luke tokenizer updated

* updated ENCODE_PLUS_ADDITIONAL_KWARGS_DOCSTRING

* aligned text and fixed puncuatations

* improved style and quality of code

* fixed error_msg in truncate_sequences

* replaced encode_plus method with regular call method

* clean up

* rephrased the docstring
      b91e65af
    • Nicolas Patry's avatar
      Enabling automatic loading of tokenizer with `pipeline` for (#13376) · c9184a2e
      Nicolas Patry authored 
      `audio-classification`.
      c9184a2e
  23. 01 Sep, 2021 4 commits