Skip to main content

Training Guide

Training documentation for reinforcement learning with Isaac Lab and behavioral cloning with LeRobot. Both frameworks run on Azure ML and NVIDIA OSMO platforms.

📖 Training Guides

GuideDescription
Isaac Lab TrainingRL training with SKRL and RSL-RL backends on Azure ML and OSMO
LeRobot TrainingBehavioral cloning with ACT and Diffusion policies
Experiment TrackingMLflow and WANDB setup, model registration, checkpoint flows
MLflow IntegrationSKRL metric logging internals, metric filtering, and troubleshooting

⚖️ Platform Comparison

AspectAzure MLOSMO
Submissionaz ml job createosmo workflow submit
OrchestrationAzure ML compute targetsOSMO workflow engine + KAI Scheduler
Experiment trackingMLflow (managed)MLflow + WANDB (credential injection)
Dataset injectionAzure ML datastoresOSMO buckets (base64 or dataset upload)
Model registrationaz ml model createVia MLflow or post-training script
MonitoringAzure ML StudioOSMO UI Dashboard

🚀 Quick Start

Isaac Lab RL training on Azure ML:

./scripts/submit-azureml-training.sh --task Isaac-Velocity-Rough-Anymal-C-v0

LeRobot behavioral cloning on OSMO:

./scripts/submit-osmo-lerobot-training.sh -d lerobot/aloha_sim_insertion_human

🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.