run 2026-10-22_v5-noflip-mask1: gate_failed

models/registry/yolo26n-seg-v5-noflip-mask1-2026-10-23/gold_report.json

@@ -0,0 +1,117 @@
+{
+  "per_class": {
+    "loop_all": {
+      "tp": 309,
+      "fp": 25,
+      "fn": 8,
+      "precision": 0.9251497005988024,
+      "recall": 0.9747634069400631,
+      "f1": 0.9493087557603687,
+      "mean_mask_iou": 0.9252735800942133,
+      "n_matched": 309
+    },
+    "pin": {
+      "tp": 214,
+      "fp": 15,
+      "fn": 10,
+      "precision": 0.9344978165938864,
+      "recall": 0.9553571428571429,
+      "f1": 0.944812362030905,
+      "mean_mask_iou": 0.9373156284625324,
+      "n_matched": 214
+    },
+    "crystal": {
+      "tp": 96,
+      "fp": 67,
+      "fn": 53,
+      "precision": 0.588957055214724,
+      "recall": 0.6442953020134228,
+      "f1": 0.6153846153846153,
+      "mean_mask_iou": 0.7870228886885559,
+      "n_matched": 96
+    },
+    "loop_face": {
+      "tp": 269,
+      "fp": 53,
+      "fn": 26,
+      "precision": 0.8354037267080745,
+      "recall": 0.911864406779661,
+      "f1": 0.8719611021069691,
+      "mean_mask_iou": 0.9008181510928498,
+      "n_matched": 269
+    },
+    "ice": {
+      "tp": 3,
+      "fp": 10,
+      "fn": 40,
+      "precision": 0.23076923076923078,
+      "recall": 0.06976744186046512,
+      "f1": 0.10714285714285714,
+      "mean_mask_iou": 0.8444813717997576,
+      "n_matched": 3
+    },
+    "needle": {
+      "tp": 0,
+      "fp": 0,
+      "fn": 1,
+      "precision": 0.0,
+      "recall": 0.0,
+      "f1": 0.0,
+      "mean_mask_iou": 0.0,
+      "n_matched": 0
+    }
+  },
+  "overall": {
+    "tp": 891,
+    "fp": 170,
+    "fn": 138,
+    "precision": 0.8397737983034873,
+    "recall": 0.8658892128279884,
+    "f1": 0.8526315789473684,
+    "mean_mask_iou": 0.9056148202172515,
+    "n_matched": 891
+  },
+  "dataset": {
+    "gold_images": 335,
+    "pred_images": 335,
+    "missing_predictions": 0,
+    "iou_threshold": 0.5
+  },
+  "environment": {
+    "host": "mx-spark-01",
+    "ultralytics": "8.4.160",
+    "torch": "2.14.0+cu130"
+  },
+  "gate": {
+    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-v4-noflip-mask1-2026-10-02/gold_report.json",
+    "f1_delta": 0.0049,
+    "mask_iou_delta": -0.0125,
+    "mask_score": 0.7842,
+    "mask_score_baseline": 0.7967,
+    "matched_only_iou_delta": -0.0063,
+    "class_f1_delta": {
+      "loop_all": -0.0133,
+      "pin": 0.0121
+    },
+    "pass_point_rule": false,
+    "rule": "paired bootstrap",
+    "bootstrap": {
+      "frames": 335,
+      "mounts": 300,
+      "draws": 2000,
+      "f1": {
+        "delta": 0.0049,
+        "lo": -0.009,
+        "hi": 0.0203,
+        "win": 0.761
+      },
+      "mask_score": {
+        "delta": -0.0125,
+        "lo": -0.0235,
+        "hi": -0.0014,
+        "win": 0.009
+      }
+    },
+    "pass": false
+  }
+}

models/registry/yolo26n-seg-v5-noflip-mask1-2026-10-23/metadata.json

@@ -0,0 +1,71 @@
+{
+  "model_id": "yolo26n-seg-v5-noflip-mask1-2026-10-23",
+  "family": "yolo26n",
+  "task": "segmentation",
+  "training_date": "2026-10-23",
+  "run": "2026-10-22_v5-noflip-mask1",
+  "dataset": "dataset/2026-10-22_v5",
+  "base_weights": "models/base/yolo26n-seg.pt",
+  "hyperparams": {
+    "epochs": 400,
+    "patience": 50,
+    "imgsz": 640,
+    "batch": 16,
+    "device": "0",
+    "overlap_mask": false,
+    "scale": 0.7,
+    "translate": 0.15,
+    "mosaic": 0.5,
+    "close_mosaic": 10,
+    "fliplr": 0,
+    "flipud": 0.3,
+    "shear": 5.0,
+    "perspective": 0.0,
+    "mask_ratio": 1
+  },
+  "export_formats": [
+    "pt"
+  ],
+  "metrics": {
+    "tp": 891,
+    "fp": 170,
+    "fn": 138,
+    "precision": 0.8397737983034873,
+    "recall": 0.8658892128279884,
+    "f1": 0.8526315789473684,
+    "mean_mask_iou": 0.9056148202172515,
+    "n_matched": 891
+  },
+  "gate": {
+    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-v4-noflip-mask1-2026-10-02/gold_report.json",
+    "f1_delta": 0.0049,
+    "mask_iou_delta": -0.0125,
+    "mask_score": 0.7842,
+    "mask_score_baseline": 0.7967,
+    "matched_only_iou_delta": -0.0063,
+    "class_f1_delta": {
+      "loop_all": -0.0133,
+      "pin": 0.0121
+    },
+    "pass_point_rule": false,
+    "rule": "paired bootstrap",
+    "bootstrap": {
+      "frames": 335,
+      "mounts": 300,
+      "draws": 2000,
+      "f1": {
+        "delta": 0.0049,
+        "lo": -0.009,
+        "hi": 0.0203,
+        "win": 0.761
+      },
+      "mask_score": {
+        "delta": -0.0125,
+        "lo": -0.0235,
+        "hi": -0.0014,
+        "win": 0.009
+      }
+    },
+    "pass": false
+  }
+}