Skip to main content

Documentation Index

Fetch the complete documentation index at: https://openpipe-art-austin-megatron-models.mintlify.app/llms.txt

Use this file to discover all available pages before exploring further.

In this Quick Start tutorial, we’ll be training Qwen3 14B Instruct to play 2048, a simple game that requires forward planning and basic math skills.
Reading time: 15 minTraining time: 2 hoursTotal cost: Free!

​
Step 1: Provision W&B API key

ART is an open source library and works across infra and observability providers. To keep things simple in this tutorial, we’ll exclusively use Weights & Biases services, which means we’ll only need to provision one API key. We’ll use these services:
  • W&B Training - autoscale GPUs for inference and training
  • W&B Models - record metrics like reward
  • W&B Weave - record your model’s traces as it generates completions
  • W&B Artifacts - store and manage your model’s checkpoints
Weights & Biases currently provides a small free tier for all the services we’ll use during this quickstart, so you shouldn’t need to add a credit card to get started. Once you have your Weights & Biases API key, open the notebook in Google Colab and set it in the Environment Variables cell. Then continue on to the next step.

​
Step 2: Run the notebook

At the top of the notebook you should see a small Run all button. Press it to begin training your model.

​
Step 3: Track metrics

While your run progresses, observe its traces and metrics in your W&B workspace. You should start seeing some progress in the first 20-30 steps. For a guide to the metrics ART logs automatically and how to add your own, see Tracking Metrics. If you have questions along the way, please ask in the Discord. Happy training!