# leftthomas/SRGAN

A PyTorch implementation of SRGAN based on CVPR 2017 paper "Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network"

Repository: https://github.com/leftthomas/SRGAN
Canonical: https://ross.abutalabs.com/products/leftthomas-srgan
Language: Python
License: MIT
License Family: permissive
Topics: pytorch, srgan, super-resolution
Last push: 2024-07-21T03:30:59+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3224, "days_push": 773, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1248, forks 353 (observed 2026-08-28T04:04:07.740108+00:00)

## What it is
A PyTorch implementation of SRGAN, the CVPR 2017 generative adversarial network for photo-realistic single-image super-resolution. It includes training, benchmark testing, and single image/video upscaling scripts with pretrained model support.

## Use cases
- upscale low resolution images with a GAN
- train an SRGAN super-resolution model on VOC2012
- evaluate super-resolution on Set5, Set14, BSD100 benchmarks
- enhance resolution of a single video
- reproduce the SRGAN paper in PyTorch

## When to choose
- you want a faithful PyTorch reproduction of the SRGAN paper
- you need 2x/4x/8x image upscaling with GAN-based detail
- you want to experiment with or fine-tune a super-resolution GAN

## When to avoid
- you need production-grade, well-maintained super-resolution (consider Real-ESRGAN)
- you want the latest transformer-based upscalers
- you need a packaged library API rather than training/testing scripts

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, image-processing, computer-vision
- domain: deep-learning, computer-vision, image-processing
- platform: python, cross-platform
- tags: pytorch, srgan, super-resolution, gan, research-code, gpu

## Member repositories
- leftthomas/SRGAN (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:07.740108+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:07:53.443849+00:00, confidence not recorded.
  - readme: https://github.com/leftthomas/SRGAN (fetched 2026-08-28T04:04:07.740108+00:00, sha df3186a77e8d)
- Data as of 2026-08-30T08:39:29.467469+00:00.