#112112th of 168 in ML frameworks and librariesFor image, video and design, image and design
retinexformer
PyTorch toolbox for low-light image enhancement with a Retinex-based transformer
- Stars
- 1.6k
- License
- MIT
- Last commit
- Nov 2026
- Language
- Python
1,428 on 21 Apr 20261,562 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Retinexformer is the ICCV 2023 code base for a transformer that brightens low-light images. It has training and testing scripts for over 15 benchmarks, including LOL, SID, SMID, SDSD and MIT Adobe FiveK. It adds DDP and mixed-precision training, self-ensemble testing, and a split-and-test mode for images up to 4000x6000. Weights are on Hugging Face and a Gradio demo runs on Spaces.
Who it is for: Researchers training or benchmarking low-light image enhancement models
Strengths
- Covers over 15 benchmarks, with pretrained weights, training logs and result downloads
- Adaptive split-and-test inference handles images up to 4000x6000
- Supports DDP, mixed-precision training and self-ensemble testing
- Second place in NTIRE 2024 low-light challenge; later winners built on it
Weaknesses
- Research code: setup needs conda, pinned PyTorch 1.11 or 2.5.1, and a BasicSR install
- Datasets are mostly on Baidu Disk and Google Drive, with manual unzip steps
- No tagged releases and no packaged library or API
- Scope is low-light enhancement only, not general image generation
What it needs
- Needs PyTorch 1.11 or 2.5.1, BasicSR, Conda, Gradio
- Models: Retinexformer, MST++, PyTorch checkpoints
Quick answers
Is retinexformer open source?
Yes. Its code is on GitHub under the MIT license.
Also in ML frameworks and libraries
See all 168| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 94 out of 100 |
| 2 | pytorchPython tensor library with GPU support and tape-based autograd | 93 out of 100 |
| 3 | ultralyticsPython library and CLI for training and running YOLO vision models | 93 out of 100 |
| #4 | pytorch-image-modelsPyTorch library of image models, pretrained weights and training scripts | 93 out of 100 |
| #5 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 92 out of 100 |