Dataset Management for Training

Manage training datasets across your infrastructure. The agent handles data staging, versioning, and distribution to training nodes.

What you'll accomplish

• Stage datasets from cloud storage to training hosts
• Track dataset versions and lineage
• Distribute data shards across nodes for distributed training
• Clean up stale datasets to free storage

Getting started

Configure your data sources in Skyportal, then ask SARA to prepare datasets for your next training run.