tao-analyze-changenet-rca
nvidiaFind the visual causes behind model failures.
model failure analysis
What it does
Investigate NVIDIA TAO Visual ChangeNet experiment failures using image evidence and produce a root-cause analysis report.
When to use it
Use it for ChangeNet model failures, poor recall, FAR or PASS-NO_PASS metrics, visual inspection audits, or AOI defect-detection RCA.
How to use it
Give it experiment results, training code, dataset files, and a target KPI; it produces a timestamped RCA report with visual evidence and analysis artifacts.
What you provide
- experiment result directory
- training code directory
- dataset directory
- target KPI
- An existing project
- Your files
Access · 2
This skill
/tmp/rca-hook-debug.log
Writes RCA hook logs
Claude Code session logs
Reads Claude Code session logs
Good to know
- starts 6 helper agents
What you need · 3
Docker must be available for the documented workflows.
NVIDIA Container Toolkit must be available for the documented workflows.
Run tao-setup first when the session was not initialized by the TAO skill bank plugin; it performs host preflight, credentials, and cross-skill discovery.
About this skill
- Visibility
- Public
- Repository
- nvidia/skills
- Created
- Oct 8, 2026
- Updated
- Oct 8, 2026
- Files
- 16