All tools
Local & on-device AIUpdated 8 days ago

Unsloth

By Unsloth AI

Open-source fine-tuning engine that trains LLMs about twice as fast with roughly 70 percent less VRAM and exports directly to GGUF and Ollama formats, with close to 69,000 GitHub stars.

Best for

  • fine-tuning open models on a single GPU
  • QLoRA and LoRA training with limited VRAM
  • no-code dataset creation and training runs
  • exporting tuned models to GGUF or Ollama
  • running and comparing models fully offline
Recent changes
  • Jul 22, 2026Added — open-source fine-tuning engine, roughly 2x faster training with 70% less VRAM, close to 69,000 GitHub stars.

Other Local & on-device AI