Fine-tune LLMs, Vision, Audio, and OCR models on your Mac
SFT, DPO, GRPO, Vision, TTS, STT, Embedding, and OCR fine-tuning — natively on MLX. Unsloth-compatible API.

Documentation · Quick Start · Training Methods

---

🖥️ Overview

MLX-Tune GUI is a pro-grade graphical orchestrator designed for researchers and developers working locally on M-series Macs. It brings a "Deep Dark" Mac-native experience to the complex world of MLX fine-tuning, allowing you to manage, monitor, and execute training runs without touching the command line.

🚀 Key Features

• Unified Training Dashboard: Orchestrate LLM (SFT/RLHF), Vision, Audio (TTS/STT), and OCR training from a single window.
• Integrated Model Library: Automatically scan and manage weights from Ollama, LM Studio, and Hugging Face local caches.
• Real-time Diagnostics: Follow every step with a built-in, resizable terminal and live log streaming.
• Persistent Settings: Save and sync your environment paths and output directories across sessions.

---

🚀 Quick Start

1. Prerequisites

Ensure you have the core library installed:

pip install mlx-tune

2. Launch the Orchestrator

The GUI is built with Next.js and interfaces directly with your local Python environment.

# Clone the repository
git clone https://github.com/Viz38/MLX-Tune-GUI.git
cd mlx-tune-gui/web

# Install and Run
npm install
npm run dev

Open http://localhost:3000 in your browser to begin training.

---

📋 v0.4.20 Orchestration Matrix

---

👥 Contributors & Credits

This is a separate, specialized repository dedicated to the MLX-Tune Graphical Suite. We acknowledge the foundational work by:

• ARahim3: Project lead and core mlx-tune architecture.
• Prince Canuma: Creator of the multi-modal MLX infrastructure.
• The MLX Team: For Apple's foundational machine learning framework.

---

Community project, not affiliated with Unsloth AI or Apple.
⭐ Star this repo if you find it useful!