b
Discover
Models
Search
About
MADTP: Multimodal Alignment-Guided Dynamic Token Pruning for Accelerating Vision-Language Transformer
6 months ago
·
CVPR