Manage end-to-end IsaacLab reinforcement learning workflows on the Euler cluster.
npx skills add https://github.com/Idate96/codex_skills --skill rl-isaaclab
Please help me install this Agent Skill. Skill: rl-isaaclab Source: https://github.com/Idate96/codex_skills/tree/main/skills/rl-isaaclab Command: npx skills add https://github.com/Idate96/codex_skills --skill rl-isaaclab
This Skill streamlines the entire Reinforcement Learning workflow within the IsaacLab environment, from local testing to cluster deployment and result synchronization.
Run a local smoke test for the excavation3d_w_cabin task with 4 environments for 3 iterations.
High-intent search queries and answers about installing and using this skill.
An end-to-end reinforcement learning workflow for IsaacLab is managed by running local smoke tests, submitting jobs to a cluster, monitoring progress, synchronizing results, and playing back policies. This ensures validation before scaling to large cluster deployments.
Validating reinforcement learning scripts before cluster submission is best done through local smoke testing. You can quickly run IsaacLab tasks on a smaller scale, such as using 4 environments for 3 iterations, to ensure scripts function correctly before committing cluster resources.
Deploying and monitoring reinforcement learning training jobs on the Euler cluster requires specific shell wrappers and environment variables. Once configured, the workflow handles job submission and tracks training progress directly on the cluster computing environment.
Trained reinforcement learning policies can be run locally for evaluation in IsaacLab using the policy playback feature. This allows you to visualize the performance of a learned policy, such as a robot arm manipulation task, after synchronizing results from the cluster.
The IsaacLab reinforcement learning workflow does require specific shell wrappers and environment variables for execution on the Euler cluster. These configurations are necessary to properly manage cluster job submission, result synchronization, and debugging operations.
gabrielfruetSubmit and monitor Slurm benchmark jobs on remote HPC clusters.
zhangyanbo2007Drive Raspberry Pi robot car via unified API for movement, vision and arm manipulation.
nebiusContainerize open-source repositories and deploy them on Nebius infrastructure.
DeepExperienceDebug and validate Relax reinforcement learning workflows on Ray clusters.
