[train] Add sampler-as-reference KL option#1263
Draft
tyler-griggs wants to merge 1 commit intomainfrom
Draft
Conversation
Allow using rollout logprobs as the KL reference instead of a separate frozen reference model, via kl_reference_source: "rollout". This eliminates the reference model memory and forward pass compute. With this option, KL constrains per-step policy drift (using the sampling policy as reference) rather than cumulative drift from initialization (using a frozen model). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
kl_reference_source: "rollout"to use rollout logprobs as KL referenceTest plan
kl_reference_source: "rollout", no ref model actors are created🤖 Generated with Claude Code