Certified Coarse-to-Fine Coordinate Refinement / registered_bench_report.json

Failed on benchmark

Raw ⬇ ZIP
  1{
  2  "bench_version": 1,
  3  "track": "two_view_gauge_localization_v2",
  4  "model": "mlp_tiny",
  5  "metric_direction": "lower is better",
  6  "n_seeds": 8,
  7  "baseline": {
  8    "best_cfg": {
  9      "lr": 0.01
 10    },
 11    "sweep": [
 12      {
 13        "cfg": {
 14          "lr": 0.001
 15        },
 16        "mean": 1.2723695039749146
 17      },
 18      {
 19        "cfg": {
 20          "lr": 0.003
 21        },
 22        "mean": 1.2335557341575623
 23      },
 24      {
 25        "cfg": {
 26          "lr": 0.01
 27        },
 28        "mean": 1.0947982370853424
 29      }
 30    ],
 31    "full": {
 32      "mean": 1.149963840842247,
 33      "std": 0.1032812344962866,
 34      "per_seed": [
 35        1.0112413167953491,
 36        1.0117237567901611,
 37        1.1340291500091553,
 38        1.222198724746704,
 39        1.1495081186294556,
 40        1.266226053237915,
 41        1.3084473609924316,
 42        1.0963362455368042
 43      ],
 44      "n": 8
 45    }
 46  },
 47  "idea": {
 48    "mean": 1.1728448644280434,
 49    "std": 0.10885410569769803,
 50    "per_seed": [
 51      0.9953804612159729,
 52      1.0588462352752686,
 53      1.2159733772277832,
 54      1.2204293012619019,
 55      1.1976563930511475,
 56      1.2475311756134033,
 57      1.355494499206543,
 58      1.0914474725723267
 59    ],
 60    "n": 8
 61  },
 62  "comparison": {
 63    "delta_mean": 0.022881023585796356,
 64    "idea_wins": 4,
 65    "n_pairs": 8,
 66    "per_seed_diffs": [
 67      -0.01586085557937622,
 68      0.04712247848510742,
 69      0.08194422721862793,
 70      -0.001769423484802246,
 71      0.048148274421691895,
 72      -0.01869487762451172,
 73      0.04704713821411133,
 74      -0.004888772964477539
 75    ],
 76    "p_value": 0.12625,
 77    "mde": 0.03145608783128045,
 78    "mde_rel_pct": 2.7353979937527115,
 79    "verdict": "no significant win",
 80    "system_worked": false
 81  },
 82  "idea_nearby_sweep": {
 83    "0.001": {
 84      "mean": 1.2723439633846283,
 85      "std": 0.07400079392237602,
 86      "per_seed": [
 87        1.1923704147338867,
 88        1.2113070487976074,
 89        1.3731907606124878,
 90        1.3125076293945312
 91      ],
 92      "n": 4
 93    },
 94    "0.01": {
 95      "mean": 1.1226573437452316,
 96      "std": 0.098156125855815,
 97      "per_seed": [
 98        0.9953804612159729,
 99        1.0588462352752686,
100        1.2159733772277832,
101        1.2204293012619019
102      ],
103      "n": 4
104    }
105  },
106  "custom_track": {
107    "name": "two_view_gauge_localization_v2",
108    "registered": true,
109    "domain": "geometry/localization"
110  },
111  "mechanism_signature": {
112    "prediction": "coarse proposal plus fixed refinement should reduce raw localization error",
113    "raw_mae_mean": 0.9011721238493919,
114    "refined_mae_mean": 0.9014216214418411,
115    "mean_proposal_shift": 0.017849167110398412,
116    "confirmed": false
117  }
118}