SCLIP: Rethinking Self-Attention for Dense Vision-Language Inference

VLM
Re-assign community