Isaac Lab is a unified and modular framework for robot learning that aims to simplify common workflows in robotics research (such as RL, learning from demonstrations, and motion planning). It is built upon NVIDIA Isaac Sim to leverage the latest simulation capabilities for photo-realistic scenes and fast and accurate simulation.

Please refer to our documentation page to learn more about the installation steps, features, tutorials, and how to set up your project with Isaac Lab.

What's New in Isaac Lab 2.0

Isaac Lab

• New Adversarial Motion Priors (AMP) examples for humanoid dancing, walking, and running
• New mimic workflow for Imitation Learning, supporting data generation to enlarge manually collected datasets
• Tiled rendering bug fixes and performance improvements, now supporting multiple tiled cameras at different resolutions
• Start up time improvements with more optimized cloning operations
• Pre-built Isaac Lab docker container hosted on NGC for headless training
• Published Isaac Lab pip package for easier installation for extension-based projects
• Selection of pre-trained policy checkpoints that can be used for inference out-of-the-box

PhysX

• Between 40% and 70% simulation throughput increases with large GPU reinforcement learning workloads involving many interactions.
• New dedicated cylinder and cone geometry designed for robot wheels that is fully GPU accelerated for reinforcement workflows. Removes limitations that previous custom geometry imposed.
• A new GPU filtering mechanism allows co-location of reinforcement learning environments at the origin with minimal performance loss for scenes with limited collider counts.
• OmniPVD: Enabled capturing reinforcement learning simulation workloads that are end-to-end GPU trained. OmniPVD viewer is limited to about 2000 environments.
• Bugfixes: Fixed performance issue for reinforcement learning GPU simulations using RigidPrimViews that have the track_contact_forces flag enabled.

License

By pulling and using the container, you accept the terms and conditions of the NVIDIA Software License Agreement.

Sources for OSS packages used in this container can be found here.

Suggested readings:

• GitHub Repo
• Documentation
• Support