mchong6/jojogan 🔢🖼️❓✓ → 🖼️

▶️ 496.1K runs 📅 Feb 2022 ⚙️ Cog 0.4.1 🔗 GitHub 📄 Paper ⚖️ License

artistic-transformation face-stylization image-stylization one-shot-learning pretrained-styles stylegan

About

JoJoGAN: One Shot Face Stylization

Example Output

Output

[object Object]

Performance Metrics

14.23s Prediction Time

60.87s Total Time

All Input Parameters

{
  "alpha": 1,
  "num_iter": 2,
  "input_face": "https://replicate.delivery/mgxm/5ef0693c-8d77-4de1-8759-52c497a33925/iu.jpeg",
  "pretrained": "sketch_multi"
}

Input Parameters

alpha Type: numberDefault: 1Range: 0 - 1: Strength of finetuned style
num_iter Type: integerDefault: 200: Number of finetuning steps (unused if pretrained style is set)
input_face (required) Type: string: Photo of human face
pretrained: Identifier of pretrained style
style_img_0 Type: string: Face style image (unused if pretrained style is set)
style_img_1 Type: string: Face style image (optional)
style_img_2 Type: string: Face style image (optional)
style_img_3 Type: string: Face style image (optional)
preserve_color Type: booleanDefault: false: Preserve the colors of the original image

Output Schema

Output

Type: string • Format: uri

Example Execution Logs

Loading e4e over the pSp framework from checkpoint: models/e4e_ffhq_encode.pt

Version Details

Version ID: 9a535b24c359049e4e7c44a976a6a54e3a1a348650084b23545fe7b55de95c2d
Version Created: September 29, 2022

Run on Replicate →