Unsloth
Unsloth Studio is a local web UI for running and fine-tuning language models. It downloads and serves GGUF, LoRA and safetensors models, compares models side by side, and exposes OpenAI- and Anthropic-compatible endpoints for tools like Claude Code and Codex.
Unsloth
In Development
This script is currently in active development and may be unstable or incomplete. Use in production environments is not recommended.
This script is currently in active development and may be unstable or incomplete. Use in production environments is not recommended.
Installation
Default install:
1
bash -c "$(curl -fsSL https://raw.githubusercontent.com/community-scripts/DevScripts/main/ct/unsloth.sh)"
Notes
Log in as unsloth with the password from ~/unsloth.creds in the container. To change it, run unsloth studio reset-password, which prints a new one.
Fine-tuning needs an NVIDIA GPU passed through to the container; without one, inference and export still work. After adding or changing a GPU, run the update to get matching builds.
Models are stored in /opt/unsloth_data and fill the disk fast (a 7B model at Q4 is about 4 GB). Add disk space if you plan to keep several.