Agent skill · nvidia
tao-analyze-gaps-od-map
Run TAO Data Services object-detection gap analysis from ground-truth and inference annotations. Use when an object detection workflow needs to identify weak images by comparing model predictions against ground truth using per-class recall, precision, and AP50 thresholds. Use when the user asks to "analyze OD gaps", "find weak OD images", or "run mAP gap analysis".
What it needs
About 4k tokens when loaded.
What this skill does
TAO Analyze Gaps OD mAP Use this skill to run TAO Data Services object-detection gap analysis. The skill compares ground-truth and inference annotations, computes per-image per-class TP/FP/FN/AP50 metrics, and identifies weak images where any class metric falls below its threshold. It does not run inference; upstream steps must produce the inference annotations first. The container entrypoint is: Inputs Required spec fields: Field Meaning --- --- groundtruthannpath KITTI label directory or COCO .json with ground-truth boxes. inferenceannpath KITTI label directory or COCO .json with model predictions. imagesdir Root image directory. Establishes the full image universe including unannotated images. resultsdir Output directory for all artifacts. kpi Identifier tag written to every output row. inputformat kitti or coco. Must be declared explicitly; never inferred from the path. Common optional fields: Field Default Meaning --- ---: --- iouthreshold 0.5 IoU at or above which a prediction is accepted as a true positive. confthreshold 0.0 Predictions below this confidence are dropped before matching. minarea 0 Boxes whose pixel area (w × h) is strictly below this value are discarded. classmapping {} Maps raw annotation label strings to canonical class names. Absent labels are kept as-is. weakthresholds {} Per-class thresholds as {classname: {recall, precision, ap50}}. Absent keys fall back to the defaultthreshold values. Reference ITS defaults: car 0.99, bicycle 0.7, person 0.7 — a strict gate on the abundant, well-learned class and looser gates on the rare ones the loop exists to improve. defaultrecallthreshold 0.5 Fallback recall threshold for classes not listed in weakthresholds. defaultprecisionthreshold 0.0 Fallback precision threshold. Set to 0.0 to disable precision-based weak selection. defaultap50threshold 0.5 Fallback for classes absent from weakthresholds. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills nvidia/tao-analyze-gaps-od-map