Workflows#
A workflow is a packaged way to run Isaac Lab somewhere other than a local Python environment. The task code, the configuration, and the command line stay the same; only the environment that executes them changes. Use this section once Isaac Lab runs locally through Installation and you need a reproducible environment, a job scheduler, or a remote GPU.
Every workflow below starts from the same container image, so they build on each other: the cluster workflow converts that image to Apptainer, and the cloud workflow provisions a machine that runs it.
Build and run Isaac Lab in a container for a reproducible environment on any Docker host. Start here – the other two workflows build on this image.
Convert the container to an Apptainer image and submit jobs through SLURM, PBS, or OSMO.
Provision a GPU workstation on AWS, GCP, Azure, or Alibaba Cloud with Isaac Automator.