[examples] SkyRL Vision Language Training - Geometry 3k#60
Open
[examples] SkyRL Vision Language Training - Geometry 3k#60
Conversation
nithinvc
commented
Apr 17, 2026
nithinvc
commented
Apr 17, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Vision language RL on geometry 3k - a math dataset.
Dataset: https://huggingface.co/datasets/hiyouga/geometry3k
Changes:
skyrl/vlm-job.yamlskyrl/Dockerfile: Base image bumped anyscale/ray:2.48.0-slim-py312-cu128 → 2.52.1-slim-py312-cu128. Shared by both jobs.skyrl/job.yaml(gsm8k): Bumped --with ray@... wheel URL from ray-2.48.0 to ray-2.52.1 to stay in sync with the new base image.skyrl/README.md: Split into shared sections (Install CLI, Clone, Understanding, View) and per-example Run sections for GSM8K and Geometry-3K VLM.The VLM example requires a local vLLM install from github and modifying SkyRL's pyproject.toml which doesn't work with the current job structure.