One Diffusion to Generate Them All

Teaser Image

This is official repo of OneDiffusion, a versatile, large-scale diffusion model that seamlessly supports bidirectional image synthesis and understanding across diverse tasks.

License

The model is trained on several non-commercially licensed datasets (e.g., DL3DV, Unsplash), thus, model weights are released under a CC BY-NC license as described in LICENSE.

Citation

@misc{le2024diffusiongenerate,
      title={One Diffusion to Generate Them All}, 
      author={Duong H. Le and Tuan Pham and Sangho Lee and Christopher Clark and Aniruddha Kembhavi and Stephan Mandt and Ranjay Krishna and Jiasen Lu},
      year={2024},
      eprint={2411.16318},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2411.16318}, 
}