← Discover
Discover · GitHub · Unsloth↑ +328 today

Train and run local AI models

Unsloth is a library that accelerates the training and execution of large language models (LLM) and diffusion models in a native environment. It makes fine-tuning processes of popular models more accessible by optimizing memory usage.

Updates

  • September 9, 2026: Stars 75,531 → 75,916, latest release v0.1.807-beta (September 8, 2026).
  • September 3, 2026: Stars 74,963 → 75,531, latest release v0.1.806-beta (September 2, 2026).
  • August 27, 2026: Stars 74,940 → 74,963, latest release v0.1.804-beta (August 27, 2026).
  • August 27, 2026: Stars 74,051 → 74,940, latest release v0.1.803-beta (August 25, 2026).

What you get

  • Run large language models and rendering tools on your local computer.
  • Perform fine-tuning processes twice as fast and with 70% less memory usage.
  • Integrate software development agents like Claude Code with your native models.

Installation

Setup for macOS, Linux and WSL
curl -fsSL https://unsloth.ai/install.sh | sh

Running it

Starting the software development agent
unsloth start claude

If you don't write code

🤖 Paste this into your AI agent (Claude Code · Codex · Antigravity)

I want to train an AI model on my local computer using Unsloth. Can you explain step by step the optimal configuration settings that will allow me to use my hardware resources in the most efficient way, optimize memory usage and speed up the fine-tuning process?

Related dictionary terms

Who it is forIt is for users who want to train and optimize their own artificial intelligence models on their local hardware and integrate them into the software projects they develop.
LicenseApache-2.0

Links

TreScout did not build this tool · we found it in GitHub trends and wrote it up. This page describes the repository as of 2026-08-14: The star count and our text belong to that day, the repository may have changed since. Check the repository link for the current state. This page was machine-translated from the Turkish original · the Turkish version prevails.