Dataset Preparation

The following sections describe the dataset preparation workflow for Imitation Learning, which consists of two main stages: Dataset Recording and Dataset Visualization. The recording stage explains how to collect data using either the Web GUI or the LeRobot CLI. The visualization stage focuses on verifying data quality by allowing users to inspect the recorded dataset.

Dataset Recording

You can record your own dataset using one of the following methods:

1. Web UI (Physical AI Manager) – Recommended

You can record your dataset using the Web UI provided by the Physical AI Manager. This option is ideal for users who prefer an intuitive interface and real-time visualization.

Option 1
Web UI (Recommended)

2. LeRobot CLI – Optional

Alternatively, you can use the LeRobot Command Line Interface (CLI) to record your dataset. This method is suitable for advanced users or automated workflows.

Option 2
LeRobot CLI (Optional)

Dataset Visualization

In the Dataset Visualization stage, you can visually inspect the collected dataset to ensure its quality.