What is QTune?
QTune is an open-source web app for fine-tuning language models on consumer GPUs. Built with Gradio, it streamlines model selection, dataset preparation, training, and deployment. It supports Hugging Face models, QLoRA configuration, real-time monitoring, and GGUF conversion, enabling accessible fine-tuning with minimal hardware requirements.
Features
- Fetch models from Hugging Face with VRAM info
- Upload or generate datasets with validation
- Configure QLoRA and training parameters
- Monitor training with real-time logs
- Convert models to GGUF for deployment
Use Cases
- Fine-tune LLMs on consumer GPUs
- Create and validate custom datasets
- Optimize models for low-VRAM hardware
- Deploy models easily with Ollama
