mlvlab / SpeaQ

Official PyTorch implementation of "Groupwise Query Specialization and Quality-Aware Multi-Assignment for Transformer-based Visual Relationship Detection" (CVPR 2024).
27 stars 2 forks source link