AutoGGUF/README.md

27 lines
910 B
Markdown

# AutoGGUF - automated GGUF model quantizer
This application provides a graphical user interface for quantizing GGUF models
using the llama.cpp library. It allows users to download different versions of
llama.cpp, manage multiple backends, and perform quantization tasks with various
options.
**Main features**:
1. Download and manage llama.cpp backends
2. Select and quantize GGUF models
3. Configure quantization parameters
4. Monitor system resources during quantization
**Usage**:
1. Install dependencies, either using the `requirements.txt` file or `pip install PyQt6 requests psutil`.
2. Run the `run.bat` script to start the application, or run the command `python src/main.py`.
**Dependencies**:
- PyQt6
- requests
- psutil
**To be implemented:**
- Actual progress bar tracking
- Download safetensors from HF and convert to unquanted GGUF
- Specify multiple KV overrides