Lightlines
By nvidia

tao-analyze-changenet-rca

nvidia1.6k installs

Find why your ChangeNet defect detector fails.

What it does

Diagnoses ChangeNet classification failures using image evidence and produces a report with quantified, prioritized fixes.

When to use it

Use to investigate ChangeNet failures, poor recall or FAR, and visual inspection pipeline quality.

How to use it

Give it experiment results, training source code, and dataset files; it returns a timestamped report with visual evidence and recommended fixes.

What you provide

  • the experiment result directory
  • the training code directory
  • the dataset directory
  • An existing project
  • Your files

Uses

Good to know

  • starts 6 helper agents
  • acts without asking

Latest review

What you need · 4

The compatibility declaration requires Docker.

The compatibility declaration requires NVIDIA Container Toolkit alongside Docker.


Access · 2

This skill

Claude Code session transcript

Read

Reads your Claude Code conversation

/tmp/rca-hook-debug.log

Write

Logs report and transcript paths