CDC Image Upscaler
Wrapped tools based on xiezw5/Component-Divide-and-Conquer-for-Real-World-Image-Super-Resolution.
First you need to install this with pip:
pip install cdc_upscalerHere is a simple example:
import logging
import os
from PIL import Image
from cdc_upscaler import image_upscale
if __name__ == '__main__':
logging.basicConfig(level=logging.INFO)
original_image = Image.open('images/your input image.png')
# any scale is supported, such as 1.5, 2, even 6 (which may take some more time)
upscaled_image = image_upscale(original_image, scale=4)
os.makedirs('output', exist_ok=True)
upscaled_image.save('output/result.png')| # | original | 4x |
|---|---|---|
| 1 |  |
 |
| 2 |  |
 |
This pretrained model is hosted on 7eu7d7/CDC_anime, which is provided by 7eu7d7. The onnx model used is hosted on narugo/CDC_anime_onnx.
