Your agent can do more. Find the skill that makes it happen.
Find your next superpower If you are an agent, refer to our llms.txt for full access.Uses denoising training and multi-scale features, with optional distillation support. Covers training, evaluation, export, distillation, quantization, and inference.
Uses the distilled bp2 commercial variant of FoundationStereo to predict disparity maps with lower latency than the full model. Supports TAO training, evaluation, export, and inference.
Predicts disparity maps from stereo image pairs. Covers TAO FoundationStereo training, evaluation, export, and inference.
Combines DINO-style detection with a BERT text encoder to detect objects described by prompts without a fixed class vocabulary. Supports training, evaluation, export, quantization, and inference.
Supports backbones including FAN, EfficientNet, and ResNet. Covers training, evaluation, distillation, quantization, export, and inference.
Learns visual representations by masking random image patches and reconstructing them. Supports pretraining, fine-tuning, evaluation, export, and inference.
Uses a ViT-MAE backbone for weakly supervised segmentation from point or bounding-box annotations. Supports TAO MAL training, evaluation, and inference.
Extends Grounding DINO with a mask-prediction head for open-set segmentation driven by text prompts. Supports training, evaluation, export, quantization, and inference.
Uses a transformer architecture with masked attention for universal image segmentation. Covers TAO training, evaluation, export, quantization, and inference.
Learns visual embeddings using triplet or contrastive losses for tasks such as retail product recognition. Supports TAO model training, evaluation, export, and inference.
Trains vision transformers without labels using teacher-student self-distillation. Supports training, export, and inference for TAO NVDINOv2 backbones that produce general-purpose visual features.
Supports training, evaluation, export, and inference for NVPanoptix3D. The model produces semantic, instance, and panoptic masks with occupancy completion using a VGGT backbone and 3D frustum reconstruction.
Skills give your agent reusable instructions for a specific job. Pick one, read what it does, and bring it into your workflow.
A name only tells half the story. Search the full description and instructions to find the right fit.
Read the skill, visit its source, and see exactly what you’re adding to your agent.
Copy the install command from a skill page and run it in your project.
npx skillycli add owner/repo --skill name