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

[Flax Encoder Decoder] Make Flax GPT2 working with cross attention #13008

Merged
merged 35 commits into from Aug 23, 2021
Merged

[Flax Encoder Decoder] Make Flax GPT2 working with cross attention #13008

merged 35 commits into from Aug 23, 2021

Commits on Aug 13, 2021

  1. make flax gpt2 working with cross attention

    @ydshieh
    ydshieh authored and ydshieh committed Aug 13, 2021
    Configuration menu
    Copy the full SHA
    d45ea91 View commit details
    Browse the repository at this point in the history

  2. Remove encoder->decoder projection layer

    @ydshieh
    ydshieh committed Aug 13, 2021
    Configuration menu
    Copy the full SHA
    d8069c7 View commit details
    Browse the repository at this point in the history

  3. A draft (incomplete) for FlaxEncoderDecoderModel

    @ydshieh
    ydshieh committed Aug 13, 2021
    Configuration menu
    Copy the full SHA
    2c68bc4 View commit details
    Browse the repository at this point in the history

  4. Add the method from_encoder_decoder_pretrained + the docstrings

    @ydshieh
    ydshieh committed Aug 13, 2021
    Configuration menu
    Copy the full SHA
    4d3c963 View commit details
    Browse the repository at this point in the history

  5. Fix the mistakes of using EncoderDecoderModel

    @ydshieh
    ydshieh committed Aug 13, 2021
    Configuration menu
    Copy the full SHA
    b71e97e View commit details
    Browse the repository at this point in the history

  6. Fix style

    @ydshieh
    ydshieh committed Aug 13, 2021
    Configuration menu
    Copy the full SHA
    02b52fc View commit details
    Browse the repository at this point in the history

  7. Add FlaxEncoderDecoderModel to the library

    @ydshieh
    ydshieh committed Aug 13, 2021
    Configuration menu
    Copy the full SHA
    d18835a View commit details
    Browse the repository at this point in the history

  8. Fix cyclic imports

    @ydshieh
    ydshieh committed Aug 13, 2021
    Configuration menu
    Copy the full SHA
    281fa2a View commit details
    Browse the repository at this point in the history

  9. Add FlaxEncoderDecoderModel to modeling_flax_auto.py

    @ydshieh
    ydshieh committed Aug 13, 2021
    Configuration menu
    Copy the full SHA
    b6f5f4f View commit details
    Browse the repository at this point in the history

  10. Remove question comments

    @ydshieh
    ydshieh committed Aug 13, 2021
    Configuration menu
    Copy the full SHA
    a36d7c3 View commit details
    Browse the repository at this point in the history

Commits on Aug 14, 2021

  1. add tests for FlaxEncoderDecoderModel

    @ydshieh
    ydshieh committed Aug 14, 2021
    Configuration menu
    Copy the full SHA
    8813d90 View commit details
    Browse the repository at this point in the history

  2. add flax_encoder_decoder to the lists of ignored entries in check_rep… 

    …o.py
    @ydshieh
    ydshieh committed Aug 14, 2021
    Configuration menu
    Copy the full SHA
    b253fee View commit details
    Browse the repository at this point in the history

  3. fix missing required positional arguments

    @ydshieh
    ydshieh committed Aug 14, 2021
    Configuration menu
    Copy the full SHA
    d0fe4b3 View commit details
    Browse the repository at this point in the history

  4. Remove **kwargs when creating FlaxEncoderDecoderModel in from_encoder… 

    …_decoder_pretrained()

Also fix generation eos/pad tokens issue
    @ydshieh
    ydshieh committed Aug 14, 2021
    Configuration menu
    Copy the full SHA
    e79d5d7 View commit details
    Browse the repository at this point in the history

  5. Fix: Use sequences from the generated_output

    @ydshieh
    ydshieh committed Aug 14, 2021
    Configuration menu
    Copy the full SHA
    27e1224 View commit details
    Browse the repository at this point in the history

  6. Change a check from assert to raise ValueError

    @ydshieh
    ydshieh committed Aug 14, 2021
    Configuration menu
    Copy the full SHA
    a3e4122 View commit details
    Browse the repository at this point in the history

  7. Fix examples and token ids issues

    @ydshieh
    ydshieh committed Aug 14, 2021
    Configuration menu
    Copy the full SHA
    3850997 View commit details
    Browse the repository at this point in the history

Commits on Aug 15, 2021

  1. Fix missing all_cross_attentions when outputting tuple in modeling_gpt2

    @ydshieh
    ydshieh committed Aug 15, 2021
    Configuration menu
    Copy the full SHA
    10de83e View commit details
    Browse the repository at this point in the history

Commits on Aug 16, 2021

  1. Remove the changes in configuration docstrings.

    @ydshieh
    ydshieh committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    70dd479 View commit details
    Browse the repository at this point in the history

  2. allow for bert 2 gpt2

    @patrickvonplaten
    patrickvonplaten committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    82ac950 View commit details
    Browse the repository at this point in the history

  3. make fix-copies

    @patrickvonplaten
    patrickvonplaten committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    f6906f0 View commit details
    Browse the repository at this point in the history

  4. Merge branch 'flax_gpt2_with_cross_attn' of https://github.com/ydshie… 

    …h/transformers into flax_gpt2_with_cross_attn
    @ydshieh
    ydshieh committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    a416698 View commit details
    Browse the repository at this point in the history

  5. Apply suggestions from code review 

    Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
    @ydshieh @patrickvonplaten
    ydshieh and patrickvonplaten committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    62a9bd7 View commit details
    Browse the repository at this point in the history

  6. Merge branch 'flax_gpt2_with_cross_attn' of https://github.com/ydshie… 

    …h/transformers into flax_gpt2_with_cross_attn
    @ydshieh
    ydshieh committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    c94754a View commit details
    Browse the repository at this point in the history

  7. Change remaining examples to bert2gpt2

    @ydshieh
    ydshieh committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    5b4849f View commit details
    Browse the repository at this point in the history

  8. Change the test to Bert2GPT2

    @ydshieh
    ydshieh committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    cf00378 View commit details
    Browse the repository at this point in the history

  9. Fix examples

    @ydshieh
    ydshieh committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    e1d5739 View commit details
    Browse the repository at this point in the history

  10. Fix import

    @ydshieh
    ydshieh committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    7410649 View commit details
    Browse the repository at this point in the history

  11. Fix unpack bug

    @ydshieh
    ydshieh committed Aug 16, 2021
    Configuration menu
    Copy the full SHA
    7f29ee5 View commit details
    Browse the repository at this point in the history

Commits on Aug 17, 2021

  1. Rename to FlaxEncoderDecoderModelTest and change the test to bert2gpt2

    @ydshieh
    ydshieh committed Aug 17, 2021
    Configuration menu
    Copy the full SHA
    fa0dde9 View commit details
    Browse the repository at this point in the history

  2. Apply suggestions from code review 

    Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
    @ydshieh @patrickvonplaten
    ydshieh and patrickvonplaten committed Aug 17, 2021
    Configuration menu
    Copy the full SHA
    109b5d1 View commit details
    Browse the repository at this point in the history

  3. Fix: NotImplentedError -> NotImplementedError

    @ydshieh
    ydshieh committed Aug 17, 2021
    Configuration menu
    Copy the full SHA
    369b3aa View commit details
    Browse the repository at this point in the history

Commits on Aug 18, 2021

  1. Apply suggestions from code review 

    Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
    @ydshieh @patrickvonplaten
    ydshieh and patrickvonplaten committed Aug 18, 2021
    Configuration menu
    Copy the full SHA
    d617f0e View commit details
    Browse the repository at this point in the history

Commits on Aug 23, 2021

  1. up

    @patrickvonplaten
    patrickvonplaten committed Aug 23, 2021
    Configuration menu
    Copy the full SHA
    0adab76 View commit details
    Browse the repository at this point in the history

  2. finalize

    @patrickvonplaten
    patrickvonplaten committed Aug 23, 2021
    Configuration menu
    Copy the full SHA
    cafc9db View commit details
    Browse the repository at this point in the history