EM-KD: Distilling Efficient Multimodal Large Language Model with Unbalanced Vision Tokens.

Ze Feng, Sen Yang, Boqiang Duan, Wankou Yang, Jingdong Wang

Browse the full AAAI paper archive.