waifu2x-python 2.0.3

High-performance Python wrapper for waifu2x-ncnn-vulkan with GPU/CPU support

Waifu2x Python

Python wrapper for waifu2x-ncnn-vulkan with a simple and powerful interface for image and video frame upscaling.

Features

• High-performance image upscaling using waifu2x-ncnn-vulkan
• Support for multiple image formats (PNG, JPG, WEBP, etc.)
• Multiple models including anime and photo styles
• GPU acceleration with fallback to CPU
• Batch processing of multiple images
• Progress tracking and error handling

Installation

pip install waifu2x-python

Dependencies

• waifu2x-ncnn-vulkan
• Python 3.8+
• OpenCV
• NumPy
• Pillow

Usage

Basic Example

from waifu2x_python import Waifu2x, ProcessingConfig

# Initialize with default settings
waifu2x = Waifu2x()

# Process a single image
result = waifu2x.process("input.jpg", "output_2x.jpg")
print(f"Processed image saved to: {result.output_path}")

Advanced Configuration

from waifu2x_python import Waifu2x, ProcessingConfig, ModelType

# Custom configuration
config = ProcessingConfig(
    model=ModelType.UPCONV_7_ANIME,  # or ModelType.CUNET, ModelType.UP_PHOTO, etc.
    scale=2,                         # 1, 2, 4, 8, 16, 32
    noise=1,                         # -1 to 3 (-1 = no effect, 0 = no noise reduction, 1-3 = noise reduction level)
    gpu_id=0,                        # -1 for CPU-only, 0 for first GPU, 1 for second GPU, etc.
    tile_size=0,                     # 0 for auto
    num_threads=4                   # Number of CPU threads to use
)

waifu2x = Waifu2x(config=config)

# Process all images in a folder
results = waifu2x.process_folder("input_folder", "output_folder")
print(f"Processed {len([r for r in results if r.success])}/{len(results)} images successfully")

Command Line Interface

# Basic usage
waifu2x -i input.jpg -o output.png

# Process folder recursively
waifu2x -i input_folder -o output_folder -r

# Custom settings
waifu2x -i input.jpg -o output.png -m models-upconv_7_anime_style_art_rgb -s 2 -n 1 -g 0 -t 200

# CPU-only mode
waifu2x -i input.jpg -o output.png -g -1

Available Models

• models-cunet - High quality denoising and upscaling (default)
• models-upconv_7_anime_style_art_rgb - For anime-style art
• models-upconv_7_photo - For photos
• models-upconv_7_anime_style_art_rgb_noise0 - For clean anime-style art
• models-upconv_7_anime_style_art_rgb_noise1 - For slightly noisy anime-style art
• models-upconv_7_anime_style_art_rgb_noise2 - For noisy anime-style art
• models-upconv_7_anime_style_art_rgb_noise3 - For very noisy anime-style art

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgements

• waifu2x-ncnn-vulkan - The underlying engine
• ncnn - High-performance neural network inference framework
• Vulkan - Cross-platform 3D graphics and compute API