Views
No views yet
[!WARNING] 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
models/raw/Pix2PixHD.| Model | Export format | Notes |
|---|---|---|
cityscapes | generator/diffusion_pytorch_model.safetensors | local-enhancer style generator checkpoint |
1Pix2PixHD-ckpt/
2 cityscapes/
3 generator/
4 config.json
5 diffusion_pytorch_model.safetensors1from PIL import Image
2
3from src.pipelines.pix2pixhd import Pix2PixHDPipeline
4
5pipe = Pix2PixHDPipeline.from_pretrained(
6 "/path/to/Pix2PixHD-ckpt/cityscapes",
7 subfolder="generator",
8 device="cuda",
9)
10
11semantic = Image.open("/path/to/semantic_input.png").convert("RGB")
12out = pipe(source_image=semantic, output_type="pil")
13out.images[0].save("pix2pixhd_output.png")generator/.model1_* modules), so it is not a strict drop-in for the current Pix2PixHDGenerator wrapper.1@inproceedings{wang2018pix2pixHD,
2 title={High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs},
3 author={Wang, Ting-Chun and Liu, Ming-Yu and Zhu, Jun-Yan and Tao, Andrew and Kautz, Jan and Catanzaro, Bryan},
4 booktitle={CVPR},
5 year={2018}
6}