Train 3D detectors from multi-camera images.
What it does
Prepares multi-camera datasets and produces trained, evaluated, exported, or quantized Sparse4D detection and tracking models.
When to use it
Use for Sparse4D adaptation and temporal perception workflows involving calibrated multi-camera images.
How to use it
Give it a Sparse4D model task and the required data or checkpoint; it prepares a specification, launches the job, and verifies its artifacts and metrics.
What you provide
- Your files
Uses
What you need · 4
Requires a compatible TAO runtime whose supported fields match the requested workflow and release.
Routes job launching through this workflow and the selected platform's job contract; whether separate installation is required is unstated.
Routes training through this skill by default when the training schema and template are available, unless the request calls for AutoML to be off; separate installation requirements are unstated.
Access · 1
This skill
TAO
Starts TAO processing jobs