Fix bugs related to loss mask, meta info, and response length#21
Merged
PeterGriffinJin merged 1 commit intoPeterGriffinJin:mainfrom Mar 19, 2025
Merged
Conversation
1. Construct the loss mask immediately after obtaining the observation to prevent encoding misalignment when converting back to tokens after text transformation. 2. Follow up on meta info to ensure that the test batch can apply do sample. 3. Remove the recording of info information for response length.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
critic/kl, information tokens should not be included, as they are not samples generated by the policy. Otherwise, it may lead to a severe negative KL explosion.After fixing these bugs, the RL training remains stable for a much longer duration:
