Find better settings for your NVIDIA TAO models.
Tunes NVIDIA TAO model actions through automated hyperparameter searches and returns the winning configuration, model artifact, and metric comparisons.
When to use it
Use for explicit TAO AutoML or hyperparameter optimization requests, including tuning training, evaluation, inference, distillation, pruning, or quantization.
Give it a TAO tuning request and the selected action's data or checkpoints; it produces experiment artifacts and identifies the winning configuration and model artifact.
What you provide
The compatibility declaration requires Docker for container-backed execution.
The compatibility declaration requires NVIDIA Container Toolkit.
This skill
nvidia-tao-automl
Installs the AutoML platform package
trial checkpoints
Deletes unneeded trial checkpoints