Skip-Vision: Efficient and Scalable Acceleration of Vision-Language Models via Adaptive Token Skipping

VLM
Re-assign community