> ## Documentation Index
> Fetch the complete documentation index at: https://mintlify.com/xinntao/Real-ESRGAN/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction

> Real-ESRGAN - Practical Algorithms for General Image/Video Restoration

<img className="block dark:hidden" src="https://raw.githubusercontent.com/xinntao/Real-ESRGAN/master/assets/realesrgan_logo.png" alt="Real-ESRGAN Logo" />

<img className="hidden dark:block" src="https://raw.githubusercontent.com/xinntao/Real-ESRGAN/master/assets/realesrgan_logo.png" alt="Real-ESRGAN Logo" />

# Real-ESRGAN

Real-ESRGAN aims at developing **Practical Algorithms for General Image/Video Restoration**. We extend the powerful ESRGAN to a practical restoration application, which is trained with pure synthetic data.

<img src="https://raw.githubusercontent.com/xinntao/Real-ESRGAN/master/assets/teaser.jpg" alt="Real-ESRGAN Comparison" />

## Key Features

<CardGroup cols={2}>
  <Card title="Multiple Models" icon="layer-group" href="#available-models">
    Specialized models for general images, anime illustrations, and anime videos
  </Card>

  <Card title="Easy Installation" icon="download" href="/installation">
    Simple pip installation with Python >= 3.7 and PyTorch >= 1.7
  </Card>

  <Card title="Quick Start" icon="rocket" href="/quickstart">
    Get started with upscaling in just a few lines of code
  </Card>

  <Card title="Face Enhancement" icon="face-smile" href="/quickstart#face-enhancement">
    Integrated GFPGAN support for enhancing faces in images
  </Card>
</CardGroup>

## Available Models

Real-ESRGAN provides several pre-trained models optimized for different use cases:

<AccordionGroup>
  <Accordion title="RealESRGAN_x4plus" icon="image" defaultOpen>
    The default model for general image super-resolution with 4x upscaling. Best for photographs and natural images.
  </Accordion>

  <Accordion title="RealESRGAN_x4plus_anime_6B" icon="wand-magic-sparkles">
    Optimized for anime illustrations with a smaller model size (6 blocks). Provides excellent results for anime-style artwork.
  </Accordion>

  <Accordion title="realesr-animevideov3" icon="film">
    Specialized model for anime videos with a compact architecture. Ideal for processing anime video frames.
  </Accordion>

  <Accordion title="realesr-general-x4v3" icon="sparkles">
    A tiny model for general scenes with denoising support. Use the `-dn` option to balance noise and avoid over-smooth results.
  </Accordion>

  <Accordion title="RealESRGAN_x2plus" icon="expand">
    2x upscaling model for cases where 4x is too aggressive.
  </Accordion>

  <Accordion title="RealESRNet_x4plus" icon="network-wired">
    Alternative 4x model with different training characteristics.
  </Accordion>
</AccordionGroup>

## Research Paper

<Note>
  Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data

  **Authors:** Xintao Wang, Liangbin Xie, Chao Dong, Ying Shan

  **Affiliation:** Tencent ARC Lab; Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences

  **Publication:** International Conference on Computer Vision Workshops (ICCVW), 2021
</Note>

[Read the Paper](https://arxiv.org/abs/2107.10833) | [Watch YouTube Video](https://www.youtube.com/watch?v=fxHWoDSSvSc) | [View Poster](https://xinntao.github.io/projects/RealESRGAN_src/RealESRGAN_poster.pdf)

## Try It Online

Before installing locally, you can try Real-ESRGAN online:

* [Replicate Demo](https://replicate.com/xinntao/realesrgan) - Web-based interface
* [Google Colab (General)](https://colab.research.google.com/drive/1k2Zod6kSHEvraybHl50Lys0LerhyTMCo?usp=sharing) - Interactive notebook for images
* [Google Colab (Anime Videos)](https://colab.research.google.com/drive/1yNl9ORUxxlL4N0keJa2SEPB61imPQd1B?usp=sharing) - Interactive notebook for anime videos

## Next Steps

<CardGroup cols={2}>
  <Card title="Installation" icon="download" href="/installation">
    Install Real-ESRGAN and its dependencies
  </Card>

  <Card title="Quick Start" icon="play" href="/quickstart">
    Run your first image upscaling example
  </Card>
</CardGroup>

## Citation

If you use Real-ESRGAN in your research, please cite:

```bibtex theme={null}
@InProceedings{wang2021realesrgan,
    author    = {Xintao Wang and Liangbin Xie and Chao Dong and Ying Shan},
    title     = {Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data},
    booktitle = {International Conference on Computer Vision Workshops (ICCVW)},
    date      = {2021}
}
```