Unsloth is an open-source, no-code deskop UI for training and running open models on your local hardware.

🦥 Unsloth Docs Unsloth is an open-source framework for running and training LLMs. Unsloth lets you run and train AI models on your own local hardware via an open-source UI. Our docs will guide you through running & training your own LLM locally.

The core Unsloth package remains licensed under Apache 2.0, while certain optional components, such as the Unsloth Studio UI are licensed under the open-source license AGPL-3.0.

No in-depth Android knowledge needed, the installer takes care of the hard parts anyway. The installer automatically cleans up conflicting GApps and only the conflicting ones because it's polite like that. 🎩 This project is available on GitHub, on GitLab, as well as on XDA.

How to connect VS Code Copilot Chat to local OpenAI-compatible models served by LM Studio or Unsloth Studio, including the BYOK utility-model setting that prevents agent errors.

This document provides a high-level overview of the user-facing interfaces of Unsloth Studio: the Command-Line Interface (CLI) and the web-based Frontend. These interfaces enable users to leverage Unsloth’s capabilities in model training, inference, exporting, and managing machine learning workflows.

In this tutorial, you will learn how to merge large language models (LLMs) easily using Unsloth Studio, a free, no-code web interface that runs entirely on your computer.

This video shows you how to load models in Unsloth Studio – and how to share them locally and across the network. Using Unsloth Studio, you can download and use models directly (much like ollama – but it appears to be much faster since it seems to use llama.cpp under the hood.

This page provides a high-level introduction to setting up and using Unsloth, including installation approaches and initial usage for both the core library and the Studio/CLI tools.

📣 Introducing Unsloth Dynamic 4-bit Quantization! We dynamically opt not to quantize certain parameters and this greatly increases accuracy while only using <10% more VRAM than BnB 4-bit.