Getting Started

How to get started with Unsloth

1

Install Unsloth

Install via pip on Linux or WSL using `pip install unsloth` or use the Docker image `unsloth/unsloth`.

2

Load Model with Hugging Face Integration

Load prequantized models and LoRA adapters using simple Python code integrated with Hugging Face tools.

3

Attach Adapters and Launch Training

Use the Hugging Face Trainer interface to attach adapters and start fine-tuning.

4

Export Fine-Tuned Model

Export the trained model to deployment formats such as GGUF, vLLM, or Hugging Face.

5

Consult Documentation

Refer to official docs for Windows setup, troubleshooting, and support for specific model types like vision or TTS.