Imitation Learning

The following workflows demonstrate end-to-end imitation learning with Isaac Lab Arena, covering teleoperation data collection, data generation, policy post-training, and closed-loop evaluation.

Currently, the following imitation learning workflow examples are provided:

• G1 Loco-Manipulation Box Pick and Place Task

• G1 Static Apple-to-Plate Task

• GR1 Open Microwave Door Task

• GR1 Sequential Pick & Place and Close Door Task

GR00T Native Environment

Arena simulation, data conversion, and evaluation clients use the standard Base container:

./docker/run_docker.sh

GR00T finetuning and GR00T policy servers use the native Isaac-GR00T uv environment from submodules/Isaac-GR00T instead of Arena’s Docker environment. Open another terminal outside the Arena Base Docker container, cd to the Isaac-GR00T checkout, and set up the environment by following the GR00T installation guide.

• G1 Loco-Manipulation Box Pick and Place Task
• Unitree G1 Static Apple-to-Plate Task
• GR1 Open Microwave Door Task
• GR1 Sequential Pick & Place and Close Door Task