BcQLM: Efficient Vision-Language Understanding with Distilled Q-Gated Cross-Modal Fusion

MLLMRe-assign community