[rl] Re-evaluate automatic resume at each chain link - #98
Merged
Conversation
penfever
enabled auto-merge (squash)
August 13, 2026 18:21
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Re-evaluate automatic checkpoint discovery when each Jupiter dependency-chain link starts. A new arm still launches with
resume_mode=nonewhen its checkpoint directory is empty; after a link banks a checkpoint, its successor resolvesresume_mode=latestand the exactglobal_step_Npath before Ray starts.The serialized job now retains whether resume was implicit AUTO or an explicit fixed choice. Explicit
none,latest,from_path, and overwrite/fresh launches keep their existing behavior.A two-link regression starts with an empty directory, writes
global_step_6, and verifies that the same serialized runner resumes it on the next invocation. The checkpoint-path and chain-guard suite passes (32 tests), and Ruff 0.15.14 passes over all tests.