[generation] automatic compilation fixes: don't log when output_attentions is True
#35989
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
gante
commented
Jan 31, 2025
|
|
||
| As a result, model components tend to be repeated across many files. There are as many attention layers defined | ||
| in `transformers` as there are models, and a significant number of those are identical to each other. | ||
| in `transformers` as there are models, and a significant number of those are identical to each other. |
Contributor
Author
There was a problem hiding this comment.
Sorry for the diff, my editor is configured to remove trailing whitespaces when saving 😅
The actual changes are L238-288
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
Contributor
Author
|
Fixing the pattern and testing its effect is exposing new bugs 🤔 Opening separate PRs to avoid bloating this one. |
This was referenced Jan 31, 2025
gante
changed the title
output_attentions is True
Contributor
Author
|
Closing -- changing this pattern might not be the best idea. Updating tests instead. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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.
What does this PR do?
WIP
[⚠️ pattern change] Skip the log regarding eager attention fallback at compilation time. Instead, when compiling, silently fallback to eager attention. This was preventing calls like
model.generate(model_inputs, cache_implementation="static", output_attentions=True, return_dict_in_generate=True)on most modelsThis fixes a bunch of failing tests and unlocks #33212