https://github.com/fscdc/rewardmap
[arxiv 2025] RewardMap: Tackling Sparse Rewards in Fine-grained Visual Reasoning via Multi-Stage Reinforcement Learning
https://github.com/fscdc/rewardmap
grpo reasoning reinforcement-learning visual-reasoning
Last synced: 7 months ago
JSON representation
[arxiv 2025] RewardMap: Tackling Sparse Rewards in Fine-grained Visual Reasoning via Multi-Stage Reinforcement Learning
- Host: GitHub
- URL: https://github.com/fscdc/rewardmap
- Owner: fscdc
- License: mit
- Created: 2025-10-01T09:02:40.000Z (8 months ago)
- Default Branch: main
- Last Pushed: 2025-10-24T08:45:09.000Z (7 months ago)
- Last Synced: 2025-10-24T10:25:41.678Z (7 months ago)
- Topics: grpo, reasoning, reinforcement-learning, visual-reasoning
- Language: Python
- Homepage: https://fscdc.github.io/RewardMap/
- Size: 147 MB
- Stars: 33
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE