Agent skills by NVIDIA, page 7

Fine-tune any HuggingFace CV / VLM / LLM model on local NVIDIA GPUs inside an NGC PyTorch container when no dedicated TAO model skill matches.

NVIDIA/skills

NV-Tesseract AD Diffusion — diffusion-based anomaly detection and fine-tuning for multivariate time series.

NVIDIA/skills

NV-Tesseract Forecasting — transformer-based multivariate time series forecasting with DARR (context-enhanced kNN retrieval), interpretability, and fine-tuning.

NVIDIA/skills

InternVideo2-CLIP L14 (TAO videoclip) for video-text retrieval, zero-shot classification, embedding extraction, LoRA fine-tuning, ONNX export, and TensorRT deployment.

NVIDIA/skills

Two-step image grounding pipeline: extracts referring expressions from (image, caption) pairs and grounds them to pixel-space bounding boxes via a VLM.

NVIDIA/skills

Four-step image referring-expression pipeline: turns images plus KITTI bounding-box labels into region descriptions, scene captions, grounded referring expressions, and (optionally) verified…

NVIDIA/skills

Multi-step video annotation pipeline that turns raw videos into Chain-of-Thought training data — multi-level captions, structured descriptions, and QA pairs (MCQ, binary, open-ended) with reasoning…

NVIDIA/skills

The mandatory pre-launch gate and four-verb execution contract for every TAO workflow or action.

NVIDIA/skills

Runs the DEFT embed-then-mine workflow for VCN AOI iterations — embeds the gap-analysis target parquet, embeds a source pool, and mines nearest-neighbour source images for downstream augmentation.

NVIDIA/skills

Run the canonical NVIDIA AOI three-phase training pipeline — Phase 1 AutoML baseline (HPO), Phase 2 DEFT loop (RCA → SDG → mining → plain-train retrain), Phase 3 AutoML refinement on the…

NVIDIA/skills

Run the full DEFT AOI improvement loop for NVIDIA TAO VisualChangeNet / ChangeNet PCB inspection models: baseline evaluate, RCA, Cosmos AnomalyGen / AMP synthetic defects, k-NN mining, retraining…

NVIDIA/skills

Start, query, and stop a network-specific TAO inference microservice ({networkarch}-inference-microservice) by delegating container execution to the appropriate platform skill.

NVIDIA/skills

Remote SLURM GPU cluster execution over SSH with sbatch/srun, Pyxis/Enroot containers, and Lustre-backed results.

NVIDIA/skills

Run a Python training/eval script directly in an existing local virtualenv — no docker, no container.

NVIDIA/skills

Action recognition from video sequences. An agent skill from NVIDIA/skills.

NVIDIA/skills

BEVFusion for multi-sensor 3D object detection. An agent skill from NVIDIA/skills.

NVIDIA/skills

CenterPose for keypoint / pose estimation. An agent skill from NVIDIA/skills.

NVIDIA/skills

Co-DETR (CoDINO) for object detection. An agent skill from NVIDIA/skills.

NVIDIA/skills

Deformable DETR for 2D object detection. An agent skill from NVIDIA/skills.

NVIDIA/skills

Monocular depth estimation using Metric Depth Anything v2 or Relative Depth Anything architectures.

NVIDIA/skills

DINO (DETR with Improved DeNoising Anchor Boxes) for 2D object detection.

NVIDIA/skills

DINOv3 continual self-supervised pre-training. An agent skill from NVIDIA/skills.

NVIDIA/skills

Real-time stereo depth estimation using FastFoundationStereo (FFS), the distilled bp2 commercial variant of FoundationStereo.

NVIDIA/skills

Stereo depth estimation using FoundationStereo. An agent skill from NVIDIA/skills.

NVIDIA/skills

Grounding DINO for open-set object detection. An agent skill from NVIDIA/skills.

NVIDIA/skills

PyTorch-based TAO image classification. An agent skill from NVIDIA/skills.

NVIDIA/skills

Masked Auto-Encoder (MAE) for self-supervised pretraining and fine-tuning.

NVIDIA/skills

MAL (Mask Auto-Label) for weakly-supervised segmentation. An agent skill from NVIDIA/skills.

NVIDIA/skills

Mask Grounding DINO for grounded instance segmentation. An agent skill from NVIDIA/skills.

NVIDIA/skills

Mask2Former for universal image segmentation (panoptic, instance, and semantic).

NVIDIA/skills

Metric-learning recognition (ml-recog) for fine-grained visual recognition.

NVIDIA/skills

NVDINOv2 for self-supervised visual representation learning.

NVIDIA/skills

NVPanoptix3D for panoptic 3D scene reconstruction from posed RGB images.

NVIDIA/skills

OCDNet for scene text detection. An agent skill from NVIDIA/skills.

NVIDIA/skills

OCRNet for scene text recognition. An agent skill from NVIDIA/skills.

NVIDIA/skills

OneFormer for universal image segmentation. An agent skill from NVIDIA/skills.

NVIDIA/skills

PointPillars for 3D object detection from LiDAR point clouds.

NVIDIA/skills

Pose classification using ST-GCN (Spatial Temporal Graph Convolutional Network).

NVIDIA/skills

Person re-identification (ReID). An agent skill from NVIDIA/skills.

NVIDIA/skills

RT-DETR (Real-Time DEtection TRansformer) for 2D object detection.

NVIDIA/skills

SegFormer for semantic segmentation. An agent skill from NVIDIA/skills.

NVIDIA/skills

Converts cuTile GPU kernels (@ct.kernel) to Triton (@triton.jit).

NVIDIA/skills

Iteratively optimize cuTile kernel performance through systematic profiling, bottleneck analysis, IR comparison, and targeted tuning.

NVIDIA/skills

Integrate TileGym kernels into Hugging Face transformers models by replacing the library's submodule(s) and certain class(es)' implementations, and patching certain class(es)' init/forward/load…

NVIDIA/skills

A skill your agent uses when deploying, operating, or integrating the VSS 3.2 GA RT-Embed Video Embedding microservice.

NVIDIA/skills

A skill your agent uses to run AutoMagicCalib on local MP4s, RTSP, or the bundled sample dataset, and to deploy vss-auto-calibration when needed.

NVIDIA/skills

A skill your agent uses to call the VIOS REST API (sensor list, timelines, clip extraction, snapshots, add/delete sensors and streams).

NVIDIA/skills

A skill your agent uses to run top-level VSS fusion search on archived video, or to ingest video files / RTSP streams for search.

NVIDIA/skills