Cast LLMEncoder output to torch.float32, freeze final layer at init.
#3900
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
arnavgarg1
approved these changes
Jan 19, 2024
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Outdated suggestions cannot be applied.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.
LLMEncoderfine-tuning is currently blocked by two issues:RuntimeError: mat1 and mat2 must have the same dtype, but got Half and Float.prepare_for_traininglogic to ECD model for LLM encoder adapter initialization #3874, the final layer of the base model is no longer frozen until training time, potentially leading to an "unused parameters" errorThis PR adds fixes for both issues:
LLMEncoderoutput is now explicitly cast totorch.float32LLMEncoder.__init__