MIT-States* dataset is a low-noise dataset in compositional zero-shot learning (CZSL), which is proposed in the paper "Warm-Cat: Warm-Started Test-Time Comprehensive Knowledge Accumulation for Compositional Zero-Shot Learning". Since original MIT-States datset suffer from substantial noise, with about 70% of its labels being incorrect, we performed a new round of annotation using INternVL-3-8b and filtration for refinement.… See the full description on the dataset page: https://huggingface.co/datasets/BJTUYXD/MIT-States_star.
