DDT: Decoupled Diffusion Transformer

Introduction

We decouple diffusion transformer into encoder-decoder design, and surpresingly that a more substantial encoder yields performance improvements as model size increases.

• We achieves 1.26 FID on ImageNet256x256 Benchmark with DDT-XL/2(22en6de).
• We achieves 1.28 FID on ImageNet512x512 Benchmark with DDT-XL/2(22en6de).
• As a byproduct, our DDT can reuse encoder among adjacent steps to accelerate inference.

Visualizations

Checkpoints

We take the off-shelf VAE to encode image into latent space, and train the decoder with DDT.

Online Demos

Coming soon.

Usages

We use ADM evaluation suite to report FID.

# for installation
pip install -r requirements.txt

# for inference
python main.py predict -c configs/repa_improved_ddt_xlen22de6_256.yaml --ckpt_path=XXX.ckpt

# for training
# extract image latent (optional)
python3 tools/cache_imlatent4.py
# train
python main.py fit -c configs/repa_improved_ddt_xlen22de6_256.yaml

Reference

@ARTICLE{ddt,
  title         = "DDT: Decoupled Diffusion Transformer",
  author        = "Wang, Shuai and Tian, Zhi and Huang, Weilin and Wang, Limin",
  month         =  apr,
  year          =  2025,
  archivePrefix = "arXiv",
  primaryClass  = "cs.CV",
  eprint        = "2504.05741"
}