Rename compute_loss in TF models #15207
Merged
Conversation
This file contains 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
2 tasks
sgugger
approved these changes
Jan 18, 2022
sgugger
approved these changes
Jan 18, 2022
There was a problem hiding this comment.
Great, let's just wait for @LysandreJik to review as well before merging.
Comment on lines
877
to
881
| logger.warning( | ||
| "The old compute_loss method is deprecated as it conflicts with the Keras compute_loss " | ||
| "method added in TF 2.8. If you want the original HF compute_loss, please call " | ||
| "hf_compute_loss() instead." | ||
| ) |
There was a problem hiding this comment.
We use warnings.warn(xxx, FutureWarning) for deprecation warning.
Also mention it will be fully removed in v5 of Transformers.
LysandreJik
approved these changes
Jan 19, 2022
Comment on lines
+872
to
+884
| def compute_loss(self, *args, **kwargs): | ||
| if hasattr(tf.keras.Model, "compute_loss"): | ||
| # This will be true in TF 2.8 or greater | ||
| return super().compute_loss(*args, **kwargs) | ||
| else: | ||
| warnings.warn( | ||
| "The old compute_loss method is deprecated as it conflicts with the Keras compute_loss " | ||
| "method added in TF 2.8. If you want the original HF compute_loss, please call " | ||
| "hf_compute_loss() instead. From TF versions >= 2.8, or Transformers versions >= 5, " | ||
| "calling compute_loss() will get the Keras method instead.", | ||
| FutureWarning, | ||
| ) | ||
| return self.hf_compute_loss(*args, **kwargs) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 renames the
compute_lossmethod on our models tohf_compute_loss, as Keras has just added acompute_lossmethod to its baseModelclass that causes lots of conflicts. Draft PR for now, since this will probably break something!