MADFormer-FFHQ / README.md
JunhaoC's picture
Upload README.md with huggingface_hub
4fef3cd verified
metadata
license: apache-2.0
datasets:
  - gaunernst/ffhq-1024-wds

MADFormer-FFHQ

This repository provides checkpoints for MADFormer trained on FFHQ-1024, combining autoregressive global conditioning and diffusion-based local refinement for high-resolution image synthesis.


πŸ“„ Paper

MADFormer: Mixed Autoregressive & Diffusion Transformers for Continuous Image Generation


πŸ“¦ Checkpoints

  • Trained for 210k steps on FFHQ-1024
  • Download checkpoint: ckpts.pt

πŸ§ͺ How to Use

# TODO

πŸ’‘ MADFormer supports flexible AR↔Diff trade-offs. On FFHQ-1024, increasing AR layer allocation yields up to 75% FID improvements under low NFE settings.


πŸ“š Citation

If you find our work useful, please cite:

@misc{chen2025madformermixedautoregressivediffusion,
      title={MADFormer: Mixed Autoregressive and Diffusion Transformers for Continuous Image Generation}, 
      author={Junhao Chen and Yulia Tsvetkov and Xiaochuang Han},
      year={2025},
      eprint={2506.07999},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2506.07999}, 
}