ACTIVE-O3: Empowering Multimodal Large Language Models with Active Perception via GRPO
aim-uofa
Official implementation of "Traceable Evidence Enhanced Visual Grounded Reasoning: Evaluation and Methodology"
Haochen-Wang409