SearchR1 reproduction fixes#65
Merged
erictang000 merged 43 commits intomainfrom Jul 10, 2025
Merged
Conversation
…m/NovaSky-AI/SkyRL into erictang000/searchr1
…m/NovaSky-AI/SkyRL into erictang000/searchr1
erictang000
commented
Jul 8, 2025
erictang000
commented
Jul 8, 2025
tyler-griggs
reviewed
Jul 8, 2025
tyler-griggs
approved these changes
Jul 9, 2025
Member
tyler-griggs
left a comment
There was a problem hiding this comment.
Approval on the assumption that you know the right way to make these updates without breaking skyrl-train's dependency on skyrl-gym :D. Does #71 resolve this concern?
Collaborator
Author
Yep! That added the same workaround I was using |
2 tasks
fannie1208
pushed a commit
to vinid/SkyRL
that referenced
this pull request
Aug 19, 2025
# What does this PR do? ### Config Change - Makes `generator.eval_sampling_params.temperature` default to 0 instead of 1. ### Adds fixes for reproducing results from SearchR1: - Updates the `skyrl-gym` `SearchEnv` implementation to correctly postprocess actions and use a consistent tool parsing schema (`<search>` and `<answer>`) - Adds session management to the `SearchToolGroup` to avoid a session being started for each request. - Updates the local retrieval server endpoint to respond to individual queries (and removed problematic torch.cuda.empty_cache() calls) - Cleaning up unnecessary printing/unused reward code paths - Adds comprehensive testing for the search environment - Adds config arguments for the search env (`skyrl_gym.search.*`) Adds docs (examples/search.rst) and (recipes.searchr1.rst) for our reproduction runs
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.
What does this PR do?
Config Change
generator.eval_sampling_params.temperaturedefault to 0 instead of 1.Adds fixes for reproducing results from SearchR1:
skyrl-gymSearchEnvimplementation to correctly postprocess actions and use a consistent tool parsing schema (<search>and<answer>)SearchToolGroupto avoid a session being started for each request.skyrl_gym.search.*)Adds docs (examples/search.rst) and (recipes.searchr1.rst) for our reproduction runs
Results below:
