| Layer | Metric | Episodes measured | Pass rate |
|---|---|---|---|
| Layer 1 · Integrity | missing/dropout | 6 | 100% |
| Layer 1 · Integrity | NaN/Inf | 6 | 100% |
| Layer 1 · Integrity | schema | 6 | 100% |
| Layer 1 · Integrity | timestamps | 6 | 100% |
| Layer 1 · Integrity | joint limits | 0 | n/a |
| Layer 1 · Integrity | video_frame_integrity | 6 | 100% |
| Layer 1 · Integrity | video_freeze | 6 | 0% |
| Layer 1 · Integrity | video_timestamp_alignment | 6 | 100% |
| Layer 1 · Integrity | video_stream_presence | 6 | 100% |
| Layer 2 · Temporal & Motion | timestamps | 0 | n/a |
| Layer 2 · Temporal & Motion | sensor sync | 0 | n/a |
| Layer 2 · Temporal & Motion | velocity/acc | 0 | n/a |
| Layer 2 · Temporal & Motion | action spikes | 0 | n/a |
| Layer 2 · Temporal & Motion | temporal_structure | 0 | n/a |
| Layer 3 · Dataset Utility | idle ratio | 0 | n/a |
| Layer 3 · Dataset Utility | coverage | 0 | n/a |
| Layer 3 · Dataset Utility | distribution | 0 | n/a |
| Metric | Finding | Evidence |
|---|---|---|
video_freeze | video_freeze: 6 episode(s) flagged (100.0%) | HARD_FAIL |
action_discontinuity | Observed action discontinuity: 108 spike(s) across 5 episode(s) (83%); statistical risk signal, not confirmed corruption | RISK_SIGNAL |
idle_ratio | Observed low effective motion: median 18.8% effective motion (81.2% idle); statistical risk signal | RISK_SIGNAL |
state_space_occupancy | State-space occupancy: median = 10.3% across episodes (6/6 episodes) | RISK_SIGNAL |
velocity_acceleration | Observed extreme acceleration spikes: 108 across 5 episode(s); statistical risk signal | RISK_SIGNAL |
REVIEW (1): 4
EXCLUDE (5): 0, 1, 2, 3, 5
REVIEW is a statistical signal worth human inspection before training — not a verdict of "bad data". EXCLUDE means an integrity-layer hard failure or unreadable episode.
D:/workbuddy-data/tmp/rda_fineart_default.json · RDA audit runs fully offline on LeRobot v2.1/v3.0 datasets · pip install robot-data-audit to reproduce · GitHub: github.com/liesliy/rda