Public Agent Skills by nvidia on Lightlines.
Guides users through CUDA-Q setup, kernel authoring, simulation targets, parallel execution, applications, and QPU access.
Guides development, testing, debugging, and contribution work on the NVIDIA cuOpt codebase.
Guides users through installing and verifying cuOpt for Python, C, or server use.
Provides configuration guidance for NeMo AutoModel launches using interactive execution, Slurm clusters, or SkyPilot cloud resources.
Routes retrieval failures to recipe families and returns concrete preview, execution, and run-report commands.
Generates synthetic CT volumes and masks through the NV-Generate-CTMR rflow-ct wrapper.
Generates synthetic brain MRI NIfTI volumes and a JSON result.
Investigate NVIDIA TAO Visual ChangeNet experiment failures using image evidence and produce a root-cause analysis report.
Submits TAO container training and inference jobs to Kubernetes clusters with NVIDIA GPU scheduling.
Trains, evaluates, exports, and runs inference on video action-recognition models using RGB, optical-flow, or joint inputs.
Trains, exports, and runs inference for a TAO NVDINOv2 backbone using self-supervised teacher-student learning.
Runs supervised TAO model training, optionally evaluates the result, and optionally exports it.