[ICCV'25] The official code of paper "Combining Similarity and Importance for Video Token Reduction on Large Visual Language Models" - View it on GitHub