Rethinking Human-Object Interaction Evaluation for both Vision-Language Models and HOI-Specific Methods

VLMRe-assign community