haopt/scflow_t2i

Official PyTorch models of "Self-Corrected Flow Distillation for Consistent One-Step and Few-Step Text-to-Image Generation" (AAAI 2025)

Quan Dao^*12†   ·   Hao Phung^*13†   ·   Trung Dao¹   ·   Dimitris N. Metaxas²   ·   Anh Tran<sup>1

1</sup>VinAI Research   ²Rutgers University   ³Cornell University

[Paper]    [Code]

^*Equal contribution   ^†Work done while at VinAI Research

Model details

We present a distilled Text-to-Image (T2I) model that supports both few-step and single-step generation. Distilled from XCLiu/instaflow_0_9B_from_sd_1_5, our model achieves an FID of 11.91 for 1-NFE generation on the COCO2014 benchmark.

Please CITE our paper and give us a :star: whenever this repository is used to help produce published results or incorporated into other software.

@inproceedings{dao2025scflow,
  title     = {Self-Corrected Flow Distillation for Consistent One-Step and Few-Step Text-to-Image Generation},
  author    = {Quan Dao and Hao Phung and Trung Dao and Dimitris Metaxas and Anh Tran},
  booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year      = {2025}
}