[EncoderDecoder] Add Cross Attention for GPT2 #6415
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
patrickvonplaten
force-pushed
the
add_gpt2_encoder_decoder
branch
from
7e7c8ad to
ad5af2c
Compare
patrickvonplaten
changed the title
sgugger
approved these changes
Aug 13, 2020
Collaborator
sgugger
left a comment
sgugger
left a comment
There was a problem hiding this comment.
Thanks, this looks great to me!
Co-authored-by: Sylvain Gugger <[email protected]>
Co-authored-by: Sylvain Gugger <[email protected]>
Codecov Report
@@ Coverage Diff @@
## master #6415 +/- ##
==========================================
- Coverage 79.98% 79.98% -0.01%
==========================================
Files 153 153
Lines 28005 28039 +34
==========================================
+ Hits 22401 22427 +26
- Misses 5604 5612 +8
Continue to review full report at Codecov.
|
sgugger
added a commit
that referenced
this pull request
Aug 14, 2020
* Generation doc * MBartForConditionalGeneration (#6441) * add MBartForConditionalGeneration * style * rebase and fixes * add mbart test in TEST_FILES_WITH_NO_COMMON_TESTS * fix docs * don't ignore mbart * doc * fix mbart fairseq link * put mbart before bart * apply doc suggestions * Use hash to clean the test dirs (#6475) * Use hash to clean the test dirs * Use hash to clean the test dirs * Use hash to clean the test dirs * fix * [EncoderDecoder] Add Cross Attention for GPT2 (#6415) * add cross attention layers for gpt2 * make gpt2 cross attention work * finish bert2gpt2 * add explicit comments * remove attention mask since not yet supported * revert attn mask in pipeline * Update src/transformers/modeling_gpt2.py Co-authored-by: Sylvain Gugger <[email protected]> * Update src/transformers/modeling_encoder_decoder.py Co-authored-by: Sylvain Gugger <[email protected]> Co-authored-by: Sylvain Gugger <[email protected]> * Sort unique_no_split_tokens to make it deterministic (#6461) * change unique_no_split_tokens's type to set * use sorted list instead of set * style * Import accuracy_score (#6480) * Apply suggestions from code review Co-authored-by: Lysandre Debut <[email protected]> * Address comments * Styling * Generation doc * Apply suggestions from code review Co-authored-by: Lysandre Debut <[email protected]> * Address comments * Styling Co-authored-by: Suraj Patil <[email protected]> Co-authored-by: Kevin Canwen Xu <[email protected]> Co-authored-by: Patrick von Platen <[email protected]> Co-authored-by: Quentin Lhoest <[email protected]> Co-authored-by: gijswijnholds <[email protected]> Co-authored-by: Lysandre Debut <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR implements Bert2GPT2 by adding cross-attention layers to GPT2.
Note that currently it is not possible to speed up decoder generation with the encoder-decoder framework (by using GPT2's past tensors) since it has to be implemented for all models that are compatible with the encoder/decoder framework (Bert, Roberta) before it can be used within the framework.
All GPT2
RUN_SLOWtests are verified to pass.Future PRs TODO: