yayafengzi/Multi-Class-OV

This is a multi-class version of open vocabulary segmentation dataset by randomly merging annotations from several classes, including ADE20K(A-150), PASCAL Context59(PC-59), and PASCAL VOC20(PAS-20). You can use them according to [📂 GitHub]. If you find this project useful in your research, please consider citing: @article{wang2025alto, title={ALTo: Adaptive-Length Tokenizer for Autoregressive Mask Generation}, author={Wang, Lingfeng and Lin, Hualing and Chen, Senda and Wang, Tao and… See the full description on the dataset page: https://huggingface.co/datasets/yayafengzi/Multi-Class-OV.