1 paper · 1 filter
Donghao Zhou, Haoyang He, Fan Zhang +10
Instruction-guided video editing has made significant progress, yet existing methods use multimodal large language models (MLLMs) primarily as semantic encoders, so they often fall…