README.md

metadata

title: Finetuner Studio
emoji: 🎛️
colorFrom: yellow
colorTo: red
sdk: gradio
sdk_version: 6.17.3
app_file: app.py
python_version: '3.12'
short_description: Low-code MLX fine-tuning studio (GUI demo)

🎛️ Finetuner Studio — GUI Demo

Low-code fine-tuning on Apple Silicon, powered by mlx-tune. This Space runs in GUI-only mode (Spaces have no Apple Silicon): explore the interface, load any Hugging Face dataset and watch the automatic format detection, and generate standalone mlx-tune training scripts with the code generator.

For actual training (12 paradigms: SFT, DPO, ORPO, SimPO, KTO, GRPO, CPT, VLM, TTS, STT, Embedding, OCR), run it on a Mac:

git clone https://github.com/aykutcayir34/finetuner
cd finetuner && pip install -e '.[mlx]' && finetuner

🇹🇷 Bu Space, Finetuner Studio arayüzünün canlı demosudur (Spaces'te Apple Silicon olmadığı için eğitim devre dışı; format algılama ve kod üretici çalışır). Örnek model: Llama-3.2-1B Turkish-Alpaca

Source: https://github.com/aykutcayir34/finetuner