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

VLM
Re-assign community