Hybrid-Level Instruction Injection for Video Token Compression in Multi-modal Large Language Models

Devs

Hybrid-Level Instruction Injection for Video Token Compression in Multi-modal Large Language Models | Read Paper on Bytez