mean_resizing = True does not work with mixed/meta initialization #3719
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
dakinggg
reviewed
Nov 20, 2024
| @@ -167,7 +167,7 @@ def _check_tokenizer_and_maybe_resize_embeddings(self, allow_embedding_resizing: | |||
| f' This would cause an error during training.' | |||
| f' Resizing the model embeddings to {len(self.tokenizer)} from {self.config.vocab_size}.', | |||
| ) | |||
| self.model.resize_token_embeddings(len(self.tokenizer)) | |||
| self.model.resize_token_embeddings(len(self.tokenizer), mean_resizing=False) | |||
There was a problem hiding this comment.
you'll need to gate on transformers version, or inspect the args to the func before passing this in or something i think
dakinggg
reviewed
Nov 20, 2024
mvpatel2000
reviewed
Nov 22, 2024
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Outdated suggestions cannot be applied.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.
What does this PR do?
Transformers recently added in
mean_resizingtoresize_token_embeddings. This is breaking with mixed initialization in downstream training tasks that requires adding tokens to Composer Huggingface Models. This PR sets this value to False for now rather than defaulting to True.