OpenClaw · Skills: nvidia

NVIDIA TAO skill for a plain train, evaluate, and export run on any TAO model.

NVIDIA TAO skill for training Sparse4D 3D detection and tracking from multiple cameras.

NVIDIA TAO skill for Visual ChangeNet pass/fail classification and change segmentation in AOI.

NVIDIA TAO skill for checking DAFT datasets for layout, schema, and reference problems.

NVIDIA TAO skill for moving a vision paper's code and recipe onto a customer dataset.

NVIDIA skill for adding a cuTile kernel operator to TileGym with tests and benchmark.

NVIDIA skill for converting cuTile Python kernels into cuTile.jl Julia versions.

NVIDIA skill for converting cuTile GPU kernels to Triton, including debugging.

NVIDIA skill for adding, changing, or debugging autotuning in cuTile kernels.

NVIDIA skill for writing high-performance GPU kernels with cuTile's tile-based model.

NVIDIA skill for iteratively speeding up cuTile kernels via profiling and tuning.

NVIDIA skill for wiring TileGym kernels into transformers model classes.

NVIDIA skill for training, evaluating, and exporting reinforcement learning policies in a Workflow.

NVIDIA skill for diagnosing and optimizing OpenUSD scene performance.

NVIDIA skill for putting a new visual question about a clip to the VSS agent.

NVIDIA skill for deploying the standalone RT-VLM captioner and calling its REST API.

NVIDIA skill for deploying and operating the 2D detection and tracking microservice for VSS.

NVIDIA skill for deploying MV3DT 3D detection with BEV fusion across calibrated cameras.

NVIDIA skill for selecting, configuring, deploying, and debugging a VSS profile.

NVIDIA skill for deploying and integrating the VSS video embedding microservice.