---
license: cc-by-nc-4.0
---
### `cadrille`: Multi-modal CAD Reconstruction with Online Reinforcement Learning

`cadrille` is trained to transform point clouds, images, and text into 3D CAD models represented as valid Python code.
`cadrille` basically is [CAD-Recode](https://huggingface.co/filapro/cad-recode-v1.5) with 2 additional modalities (images and text). 
Here, you can access the three-modal model without RL fine-tuning.
Training and testing code is available on our [github](https://github.com/col14m/cadrille).
And if you like it, give us a github 🌟.

### Citation

If you find this work useful for your research, please cite our paper:

```
@article{kolodiazhnyi2025cadrille,
  title={cadrille: Multi-modal CAD Reconstruction with Online Reinforcement Learning},
  author={Maksim Kolodiazhnyi, Denis Tarasov, Dmitrii Zhemchuzhnikov, Alexander Nikulin, Ilya Zisman, Anna Vorontsova, Anton Konushin, Vladislav Kurenkov, Danila Rukhovich},
  journal={arXiv preprint arXiv:2505.22914},
  year={2025}
}
```