we do not have a full checkpoint conversion validation, if you encounter pipeline loading failure and unsidered output, please contact me via bili_sakura@zju.edu.cn

BiliSakura/Pix2PixHD-ckpt

Packaged Pix2PixHD checkpoint from models/raw/Pix2PixHD.

Variant

Model	Export format	Notes
cityscapes	generator/diffusion_pytorch_model.safetensors	local-enhancer style generator checkpoint

Repository layout

Pix2PixHD-ckpt/
  cityscapes/
    generator/
      config.json
      diffusion_pytorch_model.safetensors

Usage (Generator Inference)

from PIL import Image

from src.pipelines.pix2pixhd import Pix2PixHDPipeline

pipe = Pix2PixHDPipeline.from_pretrained(
    "/path/to/Pix2PixHD-ckpt/cityscapes",
    subfolder="generator",
    device="cuda",
)

semantic = Image.open("/path/to/semantic_input.png").convert("RGB")
out = pipe(source_image=semantic, output_type="pil")
out.images[0].save("pix2pixhd_output.png")

Notes

• A normalized safetensors export is included under generator/.
• This checkpoint appears to be local-enhancer style (model1_* modules), so it is not a strict drop-in for the current Pix2PixHDGenerator wrapper.

Citation

@inproceedings{wang2018pix2pixHD,
  title={High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs},
  author={Wang, Ting-Chun and Liu, Ming-Yu and Zhu, Jun-Yan and Tao, Andrew and Kautz, Jan and Catanzaro, Bryan},
  booktitle={CVPR},
  year={2018}
}