Thank You & Acknowledgements
Rice Upscaler is built on cutting-edge AI research and open-source generosity. Our gratitude to these foundational projects:
๐ Project & Source
Rice Upscaler Website
Official homepage, distribution and release updates.
Rice Upscaler GitHub
Source code, issue tracker, documentation and releases.
โก AI & Upscaling Engines
DAT (Dual Aggregation Transformer)
SOTA Transformer for image super-resolution (CVPR). Preserves textures and fine details.
Real-ESRGAN
Practical image and anime restoration. Removes compression artifacts and blur.
Lanczos Resampling
High-quality windowed sinc resampling. Runs near-instantly without GPU.
๐ ๏ธ Frameworks & Libraries
Tencent ncnn
High-performance neural network inference optimized for cross-vendor Vulkan GPUs.
Microsoft DirectML
Hardware-accelerated DirectX 12 ML library for high-performance inference on Windows.
PyTorch
Leading open-source deep learning framework for AI architectures.
Pillow (PIL)
Python Imaging Library โ cornerstone of Python image processing workflows.