$γ-$MoD: Exploring Mixture-of-Depth Adaptation for Multimodal Large Language Models

VLM
MoE
Re-assign community