DragAPart

Official implementation of 'DragAPart: Learning a Part-Level Motion Prior for Articulated Objects' (ECCV 2024)

<p align="center"> [<a href="https://arxiv.org/abs/2403.15382"><strong>arXiv</strong></a>] [<a href="https://huggingface.co/spaces/rayli/DragAPart"><strong>Demo</strong></a>] [<a href="https://dragapart.github.io/"><strong>Project</strong></a>] [<a href="#citation"><strong>BibTeX</strong></a>] </p>

Inference

Please refer to the huggingface demo.

Training

accelerate launch --multi_gpu --mixed_precision fp16 --num_processes 8 train.py --config configs/train-DragAPart.yaml --wandb

Data

See the Drag-a-Move folder.

If you need the real-world images and our manually defined drags for evaluation, please contact me.

TODO

• [x] Release inference code.
• [x] Release training code.
• [x] Release dataset downloading script and dataloader code.

Citation

@article{li2024dragapart,
  title     = {DragAPart: Learning a Part-Level Motion Prior for Articulated Objects},
  author    = {Li, Ruining and Zheng, Chuanxia and Rupprecht, Christian and Vedaldi, Andrea},
  journal   = {arXiv preprint arXiv:2403.15382},
  year      = {2024}
}

Acknowledgements

We would like to thank Minghao Chen, Junyu Xie, and Laurynas Karazija for insightful discussions. This work is in part supported by a Toshiba Research Studentship and ERC-CoG UNION 101001212.