TensorZero Recipe: Supervised Fine-Tuning with Unsloth

The unsloth.ipynb notebook provides a step-by-step recipe to perform supervised fine-tuning of models using Unsloth based on data collected by the TensorZero Gateway. Set TENSORZERO_CLICKHOUSE_URL=http://chuser:chpassword@localhost:8123/tensorzero in the shell your notebook will run in.

Setup

We recommend using uv.

uv sync

Optional: Dev Container

We have provided a Dev Container config in .devcontainer to help users of VS Code who want to run the notebook on a remote server. The Dev Container pulls the Unsloth docker image. To use our container, follow the VS Code Instructions, then proceed with the "Using uv" instructions below.