|
|
--- |
|
|
license: apache-2.0 |
|
|
--- |
|
|
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 [\[📂 GitHub\]](https://github.com/yayafengzi/ALToLLM). |
|
|
|
|
|
If you find this project useful in your research, please consider citing: |
|
|
|
|
|
```BibTeX |
|
|
@article{wang2025alto, |
|
|
title={ALTo: Adaptive-Length Tokenizer for Autoregressive Mask Generation}, |
|
|
author={Wang, Lingfeng and Lin, Hualing and Chen, Senda and Wang, Tao and Cheng, Changxu and Zhong, Yangyang and Zheng, Dong and Zhao, Wuyue}, |
|
|
journal={arXiv preprint arXiv:2505.16495}, |
|
|
year={2025} |
|
|
} |
|
|
``` |
