Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

TF port of ESM #19587

Merged
merged 23 commits into from Oct 17, 2022
Merged

TF port of ESM #19587

merged 23 commits into from Oct 17, 2022

Commits on Oct 14, 2022

  1. Partial TF port for ESM model

    @Rocketknight1
    Your Name authored and Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    f77c6fd View commit details
    Browse the repository at this point in the history

  2. Add ESM-TF tests

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    8a2449b View commit details
    Browse the repository at this point in the history

  3. Add the various imports for TF-ESM

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    ddcb48d View commit details
    Browse the repository at this point in the history

  4. TF weight conversion almost ready

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    10cac9f View commit details
    Browse the repository at this point in the history

  5. Stop ignoring the decoder weights in PT

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    4f6eff2 View commit details
    Browse the repository at this point in the history

  6. Add tests and lots of fixes

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    9ce1acd View commit details
    Browse the repository at this point in the history

  7. fix-copies

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    8a89db8 View commit details
    Browse the repository at this point in the history

  8. Fix imports, add model docs

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    c68f3ef View commit details
    Browse the repository at this point in the history

  9. Add get_vocab() to tokenizer

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    f074cd7 View commit details
    Browse the repository at this point in the history

  10. Fix vocab links for pretrained files

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    840df9e View commit details
    Browse the repository at this point in the history

  11. Allow multiple inputs with a sep

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    d89bef0 View commit details
    Browse the repository at this point in the history

  12. Use EOS as SEP token because ESM vocab lacks SEP

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    345b360 View commit details
    Browse the repository at this point in the history

  13. Correctly return special tokens mask from ESM tokenizer

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    c3cc44f View commit details
    Browse the repository at this point in the history

  14. make fixup

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    5dbf24c View commit details
    Browse the repository at this point in the history

  15. Stop testing unsupported embedding resizing

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    e659a9f View commit details
    Browse the repository at this point in the history

  16. Handle TF bias correctly

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    7ebbf2d View commit details
    Browse the repository at this point in the history

  17. Skip all models with slow tokenizers in the token classification test

    @Rocketknight1
    Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    9f10249 View commit details
    Browse the repository at this point in the history

  18. Fixing the batch/unbatcher of pipelines to accomodate the None being 

    passed around.
    @Narsil @Rocketknight1
    Narsil authored and Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    357877a View commit details
    Browse the repository at this point in the history

  19. Fixing pipeline bug caused by slow tokenizer being different.

    @Narsil @Rocketknight1
    Narsil authored and Rocketknight1 committed Oct 14, 2022
    Configuration menu
    Copy the full SHA
    f5fbfb9 View commit details
    Browse the repository at this point in the history

Commits on Oct 17, 2022

  1. Update src/transformers/models/esm/modeling_tf_esm.py 

    Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>
    @Rocketknight1 @gante
    Rocketknight1 and gante committed Oct 17, 2022
    Configuration menu
    Copy the full SHA
    7f85040 View commit details
    Browse the repository at this point in the history

  2. Update src/transformers/models/esm/modeling_tf_esm.py 

    Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>
    @Rocketknight1 @gante
    Rocketknight1 and gante committed Oct 17, 2022
    Configuration menu
    Copy the full SHA
    82d39c1 View commit details
    Browse the repository at this point in the history

  3. Update src/transformers/models/esm/modeling_tf_esm.py 

    Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>
    @Rocketknight1 @gante
    Rocketknight1 and gante committed Oct 17, 2022
    Configuration menu
    Copy the full SHA
    068374f View commit details
    Browse the repository at this point in the history

  4. Update set_input_embeddings and the copyright notices

    @Rocketknight1
    Rocketknight1 committed Oct 17, 2022
    Configuration menu
    Copy the full SHA
    ca2a52f View commit details
    Browse the repository at this point in the history