This Replicate model provides an inpainting workflow for Stable Diffusion.
Image Super-Resolution
This project implements neural networks for single-image super-resolution to upscale and improve low-resolution images.
About Image Super-Resolution
Image Super-Resolution (ISR) provides Keras implementations of Residual Dense Networks for single-image super-resolution to upscale and improve low-resolution images. It includes pre-trained RDN weights (psnr-large, psnr-small, noise-cancel) and an RRDN GAN weight (gans), along with code to train models using content and adversarial loss components. Prediction and training are done in Python, with Docker scripts, Google Colab notebooks, and scripts for AWS training. Large images require patch-based prediction to avoid memory allocation errors, and the project notes compatibility with Python 3.6.
Key features
- Upscale low-resolution images with pre-trained RDN models
- Generate upscaled images with RRDN GAN-based model
- Artefact-cancelling RDN variant trained with adversarial losses
- Patch-based prediction for large images to avoid memory errors
- Train custom models with configurable content and adversarial losses
- Docker scripts and Colab notebooks for training and prediction
Best for: Python users who want to upscale low-resolution images with pre-trained models or train their own super-resolution networks.
Good for
- Four pre-trained weights available directly when creating model
- Installable from PyPI with pip install ISR
- Includes Docker scripts and Colab notebooks for training and prediction
Watch out for
- Listed as compatible with Python 3.6
- Large-image prediction can cause memory allocation errors without patch option
- Training can give patchy results and requires tuning loss weights
Image Super-Resolution at a glance
- Category
- AI & Machine Learning › AI Image Editing
- Type
- Website
- Pricing
- Not confirmed. Check the site
- Platforms
- Web
- Open source
- Not stated
- Language
- English
- Last checked
- Oct 5, 2026 · Online when checked
- Listed
- Oct 5, 2026
Image Super-Resolution alternatives
Tools and sites you could use instead, ranked by how closely they match.
Upload an image to generate a detailed text prompt that can be used to recreate it with an image model.
CodeFormer is a model for restoring old photos and AI-generated faces through a hosted playground and API.
Popular in AI Image Editing
Well-known listings from the same topic.
Kontext AIkontext-ai.comAI Photo Editoraiphotoeditor.aiDeep Imagedeep-image.aineural.loveneural.lovephotoeditor.aiphotoeditor.aiBeFunkybefunky.com
Image Super-Resolution FAQ
How do I install ISR?
Install from PyPI with pip install ISR, or clone the GitHub repository and run python setup.py install.
How do I use a pre-trained model?
Create a model with a weights option, for example RDN(weights='psnr-small') or RRDN(weights='gans'), then call predict on the image array.
How do I predict on large images?
Use the by_patch_of_size option, for example model.predict(image, by_patch_of_size=50), to avoid memory allocation errors.
Is something wrong or out of date?
Tell us and an editor will check it.
Own this website?
Claim this listing
Claim the listing and suggest updates yourself. An editor checks every claim first.