DOCSEARCHbyAlgolia
Submit Docs

QualityScaler Documentation

repository·main·Indexed 25 days ago

https://github.com/djdefrag/qualityscaler

A Windows application that uses AI and ONNX models to upscale and de-noise images and videos. Built with Python, it utilizes DirectML for DirectX 12 compatible GPUs and features a GUI, multi-GPU support, and automatic image tiling to manage VRAM limitations.

Tokens
420
Snippets
1
Records
4
Agent score
35%

What's inside QualityScaler

  1. Overview of QualityScaler

    main
    QualityScaler is a Windows-based application powered by AI designed to enhance, upscale, and de-noise both photographs and videos. It features a GUI, supports multiple GPUs, and includes automatic image tiling to prevent GPU VRAM limitations. The application is privacy-focused and can run entirely offline.

  2. Set up QualityScaler for local development

    main

    To run QualityScaler from the source code on your machine, follow these steps:

    1. Prerequisites:

    2. Installation:

      • Download the project ZIP from GitHub and extract it.
      • Extract the downloaded AI models into the /AI-onnx folder.
      • Extract ffmpeg.exe into the /Assets folder.
      • Open the project directory in VSCode.
      • Open QualityScaler.py (install Python plugins if prompted by VSCode).
      • Open the VSCode Terminal and run:
        pip install -r requirements.txt
      • Close and re-open VSCode to refresh dependencies.
      • Click the Play button in the upper right corner of VSCode to run the application.
    pip install -r requirements.txt