basicsr

Open Source Image and Video Restoration Toolbox for Super-resolution, Denoise, Deblurring, etc. Currently, it includes EDSR, RCAN, SRResNet, SRGAN, ESRGAN, EDVR, BasicVSR, SwinIR, ECBSR, etc. Also support StyleGAN2, DFDNet.

Winning Solution in NTIRE19 Challenges on Video Restoration and Enhancement (CVPR19 Workshops) - Video Restoration with Enhanced Deformable Convolutional Networks. EDVR has been merged into BasicSR and this repo is a mirror of BasicSR.

"Retinexformer: One-stage Retinex-based Transformer for Low-light Image Enhancement" (ICCV 2023 Top-10 Cited 🏆) & (NTIRE 2024 Runner-Up 🏆) & (NTIRE 2025 Winner 🏆) & (NTIRE 2026 Winner 🏆)

BasicSR-Examples illustrates how to easily use BasicSR in your own project

Official repository of the Fried Rice Lab, including code resources of the following our works: ESWT [arXiv], etc. This repository also implements many useful features and out-of-the-box image restoration models.

Windows only GUI for ESRGAN with additional features

an inference lib for image/video restoration with VapourSynth support

Fork of Basic Super-Resolution codes for development. Includes ESRGAN, SFT-GAN for training and testing.

Face Swap Finetuned Model

python notebook super resolution

an inference lib for video frame interpolation with VapourSynth support

Open Source Image and Video Restoration Toolbox for StarSRGAN.

Lightweight SR model with DFT-based frequency-gated IMDN blocks (<0.5M params), trained from scratch on DF2K and evaluated on Set5 (Y-channel PSNR, 2-pixel shave) using a BasicSR-based pipeline; targets x2/x3/x4 upscaling with improved high-frequency preservation.

A GAN architecture that improves the resolution of the input image 4 times (x4) based on "basicsr" repo

【matlab / python / tkinter / BasicSR】基于EDVR SRCNN FSRCNN 的超分辨率部署程序

Lightweight MSRResNet-based blind face super-resolution model trained on FFHQ using the BasicSR framework.

이미지 개선 딥러닝 모델 선택 적용 시스템 (NAFNet, HAT, MAXIM)

Fine-tune Real-ESRGAN on logos to upscale low-quality logos into high-res PNGs and clean SVG vectors, with a Gradio UI

Repository for R basics course

Lightweight, dependency-free terminal UI for monitoring AI training runs. Pure C, no ncurses, no Python. Pluggable profiles (BasicSR, Lightning, HuggingFace, custom) and multi-GPU backend (AMD sysfs, NVIDIA nvidia-smi).