| name: SL-HOI Weights | |
| description: Streamlined Open-Vocabulary Human-Object Interaction Detection | |
| topics: [human-object-interaction,HOI,dinov3,computer-vision,deep-learning,cvpr] | |
| license: mit | |
| # SL-HOI | |
| [Paper](https://arxiv.org/abs/2603.27500) | [GitHub](https://github.com/MPI-Lab/SL-HOI) | |
| Streamlined one-stage framework for open-vocabulary human-object interaction detection built upon DINOv3. | |
| ## Model Weights | |
| - `params/` - Pre-computed HOI classifier weights (`swig/` and `hico/`) | |
| - `pretrained/` - Trained checkpoints (`swig/`, `hico/`, `hico_ov/`) | |
| ## Citation | |
| ```bibtex | |
| @inproceedings{slhoi2026, | |
| title={Streamlined Open-Vocabulary Human-Object Interaction Detection}, | |
| author={Chang Sun and Dongliang Liao and Changxing Ding}, | |
| booktitle={CVPR}, | |
| year={2026} | |
| } | |
| ``` | |