ResCLIP: Residual Attention for Training-free Dense Vision-language Inference

VLM
Re-assign community