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
| Guide | Description |
|---|---|
| Azure ML Training | Submit Isaac Lab and LeRobot training jobs to Azure ML |
| Experiment Tracking | MLflow setup, model registration, checkpoint flows |
| Isaac Lab Training | RL training with SKRL and RSL-RL backends on Azure ML and OSMO |
| LeRobot Checkpoint Migration | Convert pre-0.6 checkpoints to the processor-based format |
| LeRobot Training | Behavioral cloning with ACT and Diffusion policies |
| MLflow Integration | SKRL metric logging internals, metric filtering, and troubleshooting |
| OSMO Training | Submit distributed Isaac Lab training jobs through NVIDIA OSMO |
⚖️ Platform Comparison
| Aspect | Azure ML | OSMO |
|---|---|---|
| Submission | az ml job create | osmo workflow submit |
| Orchestration | Azure ML compute targets | OSMO workflow engine + KAI Scheduler |
| Experiment tracking | MLflow (managed) | MLflow (Azure ML backend) |
| Dataset injection | Azure ML datastores | OSMO object storage or dataset upload |
| Model registration | az ml model create | Via MLflow or post-training script |
| Monitoring | Azure ML Studio | OSMO UI Dashboard |
🚀 Quick Start
Isaac Lab RL training on Azure ML:
training/rl/scripts/submit-azureml-training.sh --task Isaac-Velocity-Rough-Anymal-C-v0
LeRobot behavioral cloning on OSMO:
training/il/scripts/submit-osmo-lerobot-training.sh -d lerobot/aloha_sim_insertion_human
📚 Related Documentation
- Deployment Guide for infrastructure setup
- AzureML Workflows for job template reference
- OSMO Workflows for workflow template reference
- Scripts Reference for CLI usage
🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.