sczhou/codeformer
Robust face restoration algorithm for old photos / AI-generated faces
Pricing
sczhou/
codeformer
Pricing for Synexa AI models works differently from other providers. Instead of being billed by time, you are billed by input and output, making pricing more predictable.
For example, generating 100 images should cost around $0.25.
Check out our docs for more information about how per-request pricing works on Synexa.
| Provider | Price ($) | Saving (%) |
|---|---|---|
| Synexa | $0.0025 | - |
| replicate | $0.0050 | 50.0% |
Readme
CodeFormer: Robust and High-Fidelity Face Restoration for Old Photos and AI-Generated Faces
CodeFormer is a powerful and innovative face restoration algorithm specifically designed to revitalize faces in old photographs and enhance the quality of AI-generated faces. Leveraging a novel Codebook Lookup Transformer architecture, CodeFormer achieves robust blind face restoration, bringing new life to degraded or synthetic facial images. It is particularly effective in improving the visual fidelity of faces generated by models like Stable Diffusion, making it an invaluable tool for a wide range of image processing applications.
Explore More:
- Project Page: In-depth project details and comprehensive information.
- Paper: Access the academic publication for technical details and research findings.
- Video Explanation: Watch a video walkthrough and explanation of CodeFormer.
- GitHub Repository: Explore the source code, contribute, and stay updated.
- Colab Demo: Try out CodeFormer instantly in your browser.
Research Web Demo
This online demo is provided for research evaluation and demonstration purposes. For permanent use and broader applications, we recommend utilizing the officially released CodeFormer implementations. You can readily access and run CodeFormer locally via the GitHub Code or experience it through our convenient Colab Demo.
Support and Contribution
If you find CodeFormer beneficial in your research or projects, please show your support by starring our GitHub Repository. Your recognition and support are highly appreciated!
License Information
CodeFormer is released under the S-Lab License 1.0 for non-commercial research purposes. Redistribution and usage must adhere to the terms outlined in this license. Please be advised that the Replicate API for CodeFormer is restricted to non-commercial use.