tier 3 · fault_location · global view · regression · all cells measured
Headline metric mae (↓ lower is better); full metric set: mae · rmse · r2 · median_abs_error. Metrics are a frozen contract (src/evemtbench/evaluation/metrics.py); label derivation: src/evemtbench/tasks/labels.
event flt_1phg_shc · grid double_line (adapt_grid/test split) · sample 23, window #529, cubicle 3 (max-RMS) · 50 ms / 480 samples @ 9.6 kHz · top: 3-phase current, bottom: 3-phase voltage · figure provenance in assets/manifest.json
| task | grid | baseline | headline | test | benchmark | params | fit/seed | trace |
|---|---|---|---|---|---|---|---|---|
| fault_location_global | cigre_mv | majority | mae ↓ | 23.915 ± 0.000 (n=5) | 30.351 ± 0.000 (n=5) | — | 0s | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | cigre_mv | random_forest | mae ↓ | 24.131 ± 0.047 (n=5) | 30.481 ± 0.083 (n=5) | — | 12.6m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | cigre_mv | mlp | mae ↓ | 22.196 ± 0.031 (n=5) | 18.868 ± 0.943 (n=5) | 42.9M | 9.6m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | cigre_mv | gru | mae ↓ | 22.648 ± 0.381 (n=5) | 20.388 ± 2.245 (n=5) | 117k | 3.0m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | cigre_mv | cnn | mae ↓ | 22.345 ± 0.302 (n=5) | 19.521 ± 0.995 (n=5) | 124k | 3.4m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | cigre_mv | resnet | mae ↓ | 22.857 ± 0.448 (n=5) | 21.112 ± 2.293 (n=5) | 602k | 3.0m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | double_line | majority | mae ↓ | 24.416 ± 0.000 (n=5) | 31.094 ± 0.000 (n=5) | — | 0s | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | double_line | random_forest | mae ↓ | 19.776 ± 0.039 (n=5) | 17.824 ± 0.092 (n=5) | — | 4.8m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | double_line | mlp | mae ↓ | 16.338 ± 0.125 (n=5) | 6.296 ± 0.269 (n=5) | 12.0M | 3.0m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | double_line | gru | mae ↓ | 15.869 ± 0.359 (n=5) | 5.102 ± 0.715 (n=5) | 68k | 1.4m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | double_line | cnn | mae ↓ | 17.771 ± 0.093 (n=5) | 10.106 ± 0.327 (n=5) | 67k | 1.7m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | double_line | resnet | mae ↓ | 17.445 ± 0.256 (n=5) | 8.267 ± 1.416 (n=5) | 530k | 3.8m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | testgrid_110kv | majority | mae ↓ | 24.378 ± 0.000 (n=5) | 31.168 ± 0.000 (n=5) | — | 0s | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | testgrid_110kv | random_forest | mae ↓ | 21.830 ± 0.072 (n=5) | 23.947 ± 0.088 (n=5) | — | 12.6m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | testgrid_110kv | mlp | mae ↓ | 16.994 ± 0.043 (n=5) | 6.538 ± 0.115 (n=5) | 26.7M | 6.5m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | testgrid_110kv | gru | mae ↓ | 17.359 ± 0.181 (n=5) | 7.804 ± 0.638 (n=5) | 92k | 2.3m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | testgrid_110kv | cnn | mae ↓ | 18.590 ± 0.215 (n=5) | 11.585 ± 0.441 (n=5) | 94k | 3.2m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | testgrid_110kv | resnet | mae ↓ | 18.794 ± 0.250 (n=5) | 12.354 ± 0.641 (n=5) | 564k | 2.6m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | ieee39 | majority | mae ↓ | 25.224 ± 0.000 (n=5) | 31.129 ± 0.000 (n=5) | — | 0s | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | ieee39 | random_forest | mae ↓ | 24.659 ± 0.057 (n=5) | 29.816 ± 0.039 (n=5) | — | 1.3h | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | ieee39 | mlp | mae ↓ | 20.697 ± 0.115 (n=5) | 20.389 ± 0.724 (n=5) | 103.4M | 33.6m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | ieee39 | gru | mae ↓ | 23.203 ± 0.385 (n=5) | 24.926 ± 1.094 (n=5) | 211k | 9.4m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | ieee39 | cnn | mae ↓ | 21.447 ± 0.374 (n=5) | 27.133 ± 4.908 (n=5) | 234k | 12.9m | trace: output · eval-code · log · train-code · config · W&B |
| fault_location_global | ieee39 | resnet | mae ↓ | 21.816 ± 0.698 (n=5) | 21.762 ± 3.878 (n=5) | 744k | 9.1m | trace: output · eval-code · log · train-code · config · W&B |
headline & reported metrics for this view — the metric set is a frozen contract; each links to the exact function that computes it
| metric | definition | entry |
|---|---|---|
mae | mean absolute error, in the grid's own distance units | evaluate() |
rmse | root mean squared error | evaluate() |
r2 | coefficient of determination 1 − SS_res/SS_tot | evaluate() |
median_abs_error | median of absolute errors | evaluate() |
PYTHONPATH=src python -m evemtbench.baselines.runner --task fault_location_global --protocol held_out \
--grid <grid> --baseline <baseline> --seeds 0 1 2 3 4