Real-time depth estimates from stereo images.
What it does
Trains, evaluates, and deploys FastFoundationStereo models that estimate depth from paired images.
When to use it
Use when training, evaluating, exporting, or running inference for a TAO FastFoundationStereo model.
How to use it
Give it stereo image data and the checkpoint needed for the chosen workflow; it produces model checkpoints, disparity predictions, or deployment files.
What you provide
- Your files
What you need · 4
Requires Docker to launch the model container.
Requires NVIDIA Container Toolkit alongside Docker.
For a standalone installation, run this skill first for host preflight, credentials, and cross-skill discovery; this prerequisite is conditional on the session not having been initialized by the TAO skill bank plugin.
Routes training here by default when AutoML is on and the train schema and template are packaged. Direct training is allowed when AutoML is off or those files are missing, so this dependency is conditional.
Access · 0
No additional actions listed.