Scaling Video-Language Models to 10K Frames via Hierarchical Differential Distillation.

Chuanqi Cheng, Jian Guan, Wei Wu, Rui Yan

Browse the full ICML paper archive.