RLLaVA
Collection
An RL-central Framework for Language and Vision Assistant, which decouples algorithm logic from distributed execution, enables modular customization.
•
4 items
•
Updated
Base model
Qwen/Qwen2.5-VL-3B-Instruct