From 3d39d405608b7de0ebfa20fb77d7ba851aa9d395 Mon Sep 17 00:00:00 2001
From: jungfrau <jungfrau@mx-spark-01.psi.ch>
Date: Tue, 29 Sep 2026 17:57:18 +0200
Subject: [PATCH] run 2026-09-29_v2-recipeA: gate_failed; comparison updated

---
 .../gold_report.json                          |  93 +++++
 .../metadata.json                             |  52 +++
 .../weights.pt                                |   3 +
 runs/2026-09-29_v2-recipeA/metrics.json       |  38 ++
 runs/2026-09-29_v2-recipeA/run.json           |  52 ++-
 runs/README.md                                |  13 +
 runs/clean_gold_comparison.json               | 346 +++++++++++-------
 7 files changed, 454 insertions(+), 143 deletions(-)
 create mode 100644 models/registry/yolo26n-seg-v2-recipeA-2026-09-29/gold_report.json
 create mode 100644 models/registry/yolo26n-seg-v2-recipeA-2026-09-29/metadata.json
 create mode 100644 models/registry/yolo26n-seg-v2-recipeA-2026-09-29/weights.pt
 create mode 100644 runs/2026-09-29_v2-recipeA/metrics.json

diff --git a/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/gold_report.json b/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/gold_report.json
new file mode 100644
index 0000000..135d8b8
--- /dev/null
+++ b/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/gold_report.json
@@ -0,0 +1,93 @@
+{
+  "per_class": {
+    "loop_all": {
+      "tp": 187,
+      "fp": 25,
+      "fn": 5,
+      "precision": 0.8820754716981132,
+      "recall": 0.9739583333333334,
+      "f1": 0.9257425742574258,
+      "mean_mask_iou": 0.8721341451334437,
+      "n_matched": 187
+    },
+    "pin": {
+      "tp": 136,
+      "fp": 21,
+      "fn": 10,
+      "precision": 0.8662420382165605,
+      "recall": 0.9315068493150684,
+      "f1": 0.8976897689768977,
+      "mean_mask_iou": 0.9176566673708884,
+      "n_matched": 136
+    },
+    "crystal": {
+      "tp": 37,
+      "fp": 33,
+      "fn": 29,
+      "precision": 0.5285714285714286,
+      "recall": 0.5606060606060606,
+      "f1": 0.5441176470588236,
+      "mean_mask_iou": 0.7928724346781946,
+      "n_matched": 37
+    },
+    "loop_face": {
+      "tp": 147,
+      "fp": 47,
+      "fn": 25,
+      "precision": 0.7577319587628866,
+      "recall": 0.8546511627906976,
+      "f1": 0.8032786885245901,
+      "mean_mask_iou": 0.8628361421194072,
+      "n_matched": 147
+    },
+    "ice": {
+      "tp": 8,
+      "fp": 35,
+      "fn": 32,
+      "precision": 0.18604651162790697,
+      "recall": 0.2,
+      "f1": 0.19277108433734938,
+      "mean_mask_iou": 0.7778219649966538,
+      "n_matched": 8
+    },
+    "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": 515,
+    "fp": 161,
+    "fn": 102,
+    "precision": 0.7618343195266272,
+    "recall": 0.8346839546191248,
+    "f1": 0.7965970610982211,
+    "mean_mask_iou": 0.8743420594116778,
+    "n_matched": 515
+  },
+  "dataset": {
+    "gold_images": 203,
+    "pred_images": 203,
+    "missing_predictions": 0,
+    "iou_threshold": 0.5
+  },
+  "gate": {
+    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json",
+    "f1_delta": -0.0123,
+    "mask_iou_delta": -0.0039,
+    "mask_score": 0.7298,
+    "mask_score_baseline": 0.7337,
+    "matched_only_iou_delta": -0.0168,
+    "class_f1_delta": {
+      "loop_all": 0.0089,
+      "pin": 0.0021
+    },
+    "pass": false
+  }
+}
diff --git a/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/metadata.json b/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/metadata.json
new file mode 100644
index 0000000..8451076
--- /dev/null
+++ b/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/metadata.json
@@ -0,0 +1,52 @@
+{
+  "model_id": "yolo26n-seg-v2-recipeA-2026-09-29",
+  "family": "yolo26n",
+  "task": "segmentation",
+  "training_date": "2026-09-29",
+  "run": "2026-09-29_v2-recipeA",
+  "dataset": "dataset/2026-09-29_v2",
+  "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": 0.5,
+    "flipud": 0.3,
+    "shear": 5.0,
+    "perspective": 0.0,
+    "mask_ratio": 4
+  },
+  "export_formats": [
+    "pt"
+  ],
+  "metrics": {
+    "tp": 515,
+    "fp": 161,
+    "fn": 102,
+    "precision": 0.7618343195266272,
+    "recall": 0.8346839546191248,
+    "f1": 0.7965970610982211,
+    "mean_mask_iou": 0.8743420594116778,
+    "n_matched": 515
+  },
+  "gate": {
+    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json",
+    "f1_delta": -0.0123,
+    "mask_iou_delta": -0.0039,
+    "mask_score": 0.7298,
+    "mask_score_baseline": 0.7337,
+    "matched_only_iou_delta": -0.0168,
+    "class_f1_delta": {
+      "loop_all": 0.0089,
+      "pin": 0.0021
+    },
+    "pass": false
+  }
+}
diff --git a/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/weights.pt b/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/weights.pt
new file mode 100644
index 0000000..a66710a
--- /dev/null
+++ b/models/registry/yolo26n-seg-v2-recipeA-2026-09-29/weights.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6cc137390a1ab1c8b90f1edf9000f37cdb6e966a47a5fdbc2b97c09d65c1fcfd
+size 6567069
diff --git a/runs/2026-09-29_v2-recipeA/metrics.json b/runs/2026-09-29_v2-recipeA/metrics.json
new file mode 100644
index 0000000..c44101a
--- /dev/null
+++ b/runs/2026-09-29_v2-recipeA/metrics.json
@@ -0,0 +1,38 @@
+{
+  "overall": {
+    "tp": 515,
+    "fp": 161,
+    "fn": 102,
+    "precision": 0.7618343195266272,
+    "recall": 0.8346839546191248,
+    "f1": 0.7965970610982211,
+    "mean_mask_iou": 0.8743420594116778,
+    "n_matched": 515
+  },
+  "gate": {
+    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json",
+    "f1_delta": -0.0123,
+    "mask_iou_delta": -0.0039,
+    "mask_score": 0.7298,
+    "mask_score_baseline": 0.7337,
+    "matched_only_iou_delta": -0.0168,
+    "class_f1_delta": {
+      "loop_all": 0.0089,
+      "pin": 0.0021
+    },
+    "pass": false
+  },
+  "training": {
+    "epochs_run": 159,
+    "precision(B)": 0.66802,
+    "recall(B)": 0.71022,
+    "mAP50(B)": 0.68898,
+    "mAP50-95(B)": 0.52278,
+    "precision(M)": 0.66354,
+    "recall(M)": 0.71012,
+    "mAP50(M)": 0.68771,
+    "mAP50-95(M)": 0.50363
+  },
+  "version": "yolo26n-seg-v2-recipeA-2026-09-29",
+  "measured_at": "2026-09-29T17:56:27"
+}
diff --git a/runs/2026-09-29_v2-recipeA/run.json b/runs/2026-09-29_v2-recipeA/run.json
index 100c4c7..c214814 100644
--- a/runs/2026-09-29_v2-recipeA/run.json
+++ b/runs/2026-09-29_v2-recipeA/run.json
@@ -2,18 +2,60 @@
   "id": "2026-09-29_v2-recipeA",
   "dataset": "dataset/2026-09-29_v2",
   "base_weights": "models/base/yolo26n-seg.pt",
-  "status": "exported",
+  "status": "gate_failed",
   "created": "2026-09-29T17:06:19",
   "created_by": "jungfrau@mx-spark-01",
-  "hyperparams": {},
-  "version": null,
-  "metrics": {},
+  "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": 0.5,
+    "flipud": 0.3,
+    "shear": 5.0,
+    "perspective": 0.0,
+    "mask_ratio": 4
+  },
+  "version": "yolo26n-seg-v2-recipeA-2026-09-29",
+  "metrics": {
+    "f1": 0.7965970610982211,
+    "mean_mask_iou": 0.8743420594116778,
+    "gate_pass": false,
+    "baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json",
+    "measured_at": "2026-09-29T17:56:27"
+  },
   "history": [
     {
       "status": "exported",
       "at": "2026-09-29T17:06:19",
       "by": "jungfrau@mx-spark-01",
       "note": "settings A (fliplr 0.5, mask_ratio 4) on dataset v2, Spark"
+    },
+    {
+      "status": "training",
+      "at": "2026-09-29T17:06:19",
+      "by": "jungfrau@mx-spark-01",
+      "note": "train_run.py"
+    },
+    {
+      "status": "measured",
+      "at": "2026-09-29T17:56:27",
+      "by": "jungfrau@mx-spark-01",
+      "note": ""
+    },
+    {
+      "status": "gate_failed",
+      "at": "2026-09-29T17:56:27",
+      "by": "jungfrau@mx-spark-01",
+      "note": "{\"baseline\": \"/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json\", \"f1_delta\": -0.0123, \"mask_iou_delta\": -0.0039, \"mask_score\": 0.7298, \"mask_score_baseline\": 0.7337, \"matched_only_iou_delta\": -0.0168, \"class_f1_delta\": {\"loop_all\": 0.0089, \"pin\": 0.0021}, \"pass\": false}"
     }
-  ]
+  ],
+  "host": "mx-spark-01",
+  "started": "2026-09-29T17:06:19"
 }
diff --git a/runs/README.md b/runs/README.md
index ab34aa1..73a0312 100644
--- a/runs/README.md
+++ b/runs/README.md
@@ -11,6 +11,8 @@ Generated by `python tools/ledger.py render`; do not edit by hand.
 | 2026-09-28_alc-v1-noflip | gate_passed | dataset/2026-09-28_alc-v1 | /home/jungfrau/AareLC/models/yolo26n-seg.pt | yolo26n-seg-alc-v1-noflip-2026-09-28 | 0.835 | 0.881 | yes | 2026-09-28T21:30 |
 | 2026-09-28_alc-v1-overlap | aborted | dataset/2026-09-28_alc-v1 | /home/jungfrau/AareLC/models/yolo26n-seg.pt | - | - | - | - | 2026-09-28T23:55 |
 | 2026-09-29_alc-v1-noflip-mask1 | promoted | dataset/2026-09-28_alc-v1 | /home/jungfrau/AareLC/models/yolo26n-seg.pt | yolo26n-seg-alc-v1-noflip-mask1-2026-09-29 | 0.827 | 0.894 | yes | 2026-09-29T11:13 |
+| 2026-09-29_v2-noflip-mask1 | exported | dataset/2026-09-29_v2 | models/base/yolo26n-seg.pt | - | - | - | - | 2026-09-29T17:06 |
+| 2026-09-29_v2-recipeA | gate_failed | dataset/2026-09-29_v2 | models/base/yolo26n-seg.pt | yolo26n-seg-v2-recipeA-2026-09-29 | 0.797 | 0.874 | no | 2026-09-29T17:56 |
 
 States: exported → queued → training → measured → gate_passed → gate_failed → promoted → deployed (or aborted).
 
@@ -67,3 +69,14 @@ States: exported → queued → training → measured → gate_passed → gate_f
 - 2026-09-29T10:36:12 · measured · jungfrau@mx-spark-01
 - 2026-09-29T10:36:12 · gate_passed · jungfrau@mx-spark-01 — {"baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-2026-09-28/gold_report.json", "f1_delta": 0.0062, "mask_iou_delta": 0.0249, "mask_score": 0.7352, "mask_score_baseline": 0.7103, "matched_only_iou_delta": 0.0133, "class_f1_delta": {"loop_all": 0.0084, "pin": -0.0167}, "pass": true}
 - 2026-09-29T11:13:44 · promoted · jungfrau@mx-spark-01 — by decision after the A/B comparison on the unseen frames (gate had passed vs A)
+
+## 2026-09-29_v2-noflip-mask1
+
+- 2026-09-29T17:06:19 · exported · jungfrau@mx-spark-01 — default recipe (no flip, full-resolution masks) on dataset v2: to be trained on the cluster with sbatch slurm/train.sbatch
+
+## 2026-09-29_v2-recipeA
+
+- 2026-09-29T17:06:19 · exported · jungfrau@mx-spark-01 — settings A (fliplr 0.5, mask_ratio 4) on dataset v2, Spark
+- 2026-09-29T17:06:19 · training · jungfrau@mx-spark-01 — train_run.py
+- 2026-09-29T17:56:27 · measured · jungfrau@mx-spark-01
+- 2026-09-29T17:56:27 · gate_failed · jungfrau@mx-spark-01 — {"baseline": "/home/jungfrau/AareLC_train/models/registry/yolo26n-seg-alc-v1-noflip-mask1-2026-09-29/gold_report.json", "f1_delta": -0.0123, "mask_iou_delta": -0.0039, "mask_score": 0.7298, "mask_score_baseline": 0.7337, "matched_only_iou_delta": -0.0168, "class_f1_delta": {"loop_all": 0.0089, "pin": 0.0021}, "pass": false}
diff --git a/runs/clean_gold_comparison.json b/runs/clean_gold_comparison.json
index 2262760..bebdee5 100644
--- a/runs/clean_gold_comparison.json
+++ b/runs/clean_gold_comparison.json
@@ -1,55 +1,55 @@
 {
-  "clean_frames": 66,
+  "clean_frames": 149,
   "reports": {
     "beamline yolo26n-seg-overlap-false-2026-04-12": {
       "per_class": {
         "loop_all": {
           "tp": 43,
           "fp": 14,
-          "fn": 21,
+          "fn": 96,
           "precision": 0.7543859649122807,
-          "recall": 0.671875,
-          "f1": 0.7107438016528924,
+          "recall": 0.30935251798561153,
+          "f1": 0.4387755102040817,
           "mean_mask_iou": 0.8603647813074378,
           "n_matched": 43
         },
         "pin": {
           "tp": 48,
           "fp": 7,
-          "fn": 11,
+          "fn": 68,
           "precision": 0.8727272727272727,
-          "recall": 0.8135593220338984,
-          "f1": 0.8421052631578948,
+          "recall": 0.41379310344827586,
+          "f1": 0.5614035087719298,
           "mean_mask_iou": 0.9478639299321778,
           "n_matched": 48
         },
         "crystal": {
           "tp": 2,
           "fp": 15,
-          "fn": 5,
+          "fn": 38,
           "precision": 0.11764705882352941,
-          "recall": 0.2857142857142857,
-          "f1": 0.16666666666666666,
+          "recall": 0.05,
+          "f1": 0.07017543859649122,
           "mean_mask_iou": 0.8492457223853397,
           "n_matched": 2
         },
         "loop_face": {
           "tp": 30,
           "fp": 12,
-          "fn": 23,
+          "fn": 90,
           "precision": 0.7142857142857143,
-          "recall": 0.5660377358490566,
-          "f1": 0.631578947368421,
+          "recall": 0.25,
+          "f1": 0.37037037037037035,
           "mean_mask_iou": 0.8979746795878049,
           "n_matched": 30
         },
         "ice": {
           "tp": 1,
           "fp": 4,
-          "fn": 1,
+          "fn": 9,
           "precision": 0.2,
-          "recall": 0.5,
-          "f1": 0.28571428571428575,
+          "recall": 0.1,
+          "f1": 0.13333333333333333,
           "mean_mask_iou": 0.8931750741839762,
           "n_matched": 1
         }
@@ -57,17 +57,17 @@
       "overall": {
         "tp": 124,
         "fp": 52,
-        "fn": 61,
+        "fn": 301,
         "precision": 0.7045454545454546,
-        "recall": 0.6702702702702703,
-        "f1": 0.6869806094182825,
+        "recall": 0.2917647058823529,
+        "f1": 0.41264559068219636,
         "mean_mask_iou": 0.9034198478996224,
         "n_matched": 124
       },
       "dataset": {
-        "gold_images": 66,
+        "gold_images": 149,
         "pred_images": 120,
-        "missing_predictions": 0,
+        "missing_predictions": 83,
         "iou_threshold": 0.5
       }
     },
@@ -76,47 +76,47 @@
         "loop_all": {
           "tp": 59,
           "fp": 10,
-          "fn": 5,
+          "fn": 80,
           "precision": 0.855072463768116,
-          "recall": 0.921875,
-          "f1": 0.887218045112782,
+          "recall": 0.4244604316546763,
+          "f1": 0.5673076923076923,
           "mean_mask_iou": 0.8815339593847193,
           "n_matched": 59
         },
         "pin": {
           "tp": 56,
           "fp": 5,
-          "fn": 3,
+          "fn": 60,
           "precision": 0.9180327868852459,
-          "recall": 0.9491525423728814,
-          "f1": 0.9333333333333333,
+          "recall": 0.4827586206896552,
+          "f1": 0.632768361581921,
           "mean_mask_iou": 0.9197473537826396,
           "n_matched": 56
         },
         "crystal": {
           "tp": 4,
           "fp": 2,
-          "fn": 3,
+          "fn": 36,
           "precision": 0.6666666666666666,
-          "recall": 0.5714285714285714,
-          "f1": 0.6153846153846153,
+          "recall": 0.1,
+          "f1": 0.1739130434782609,
           "mean_mask_iou": 0.8425290211607122,
           "n_matched": 4
         },
         "loop_face": {
           "tp": 30,
           "fp": 11,
-          "fn": 23,
+          "fn": 90,
           "precision": 0.7317073170731707,
-          "recall": 0.5660377358490566,
-          "f1": 0.6382978723404255,
+          "recall": 0.25,
+          "f1": 0.37267080745341613,
           "mean_mask_iou": 0.8851212986608521,
           "n_matched": 30
         },
         "ice": {
           "tp": 0,
           "fp": 0,
-          "fn": 2,
+          "fn": 10,
           "precision": 0.0,
           "recall": 0.0,
           "f1": 0.0,
@@ -127,17 +127,17 @@
       "overall": {
         "tp": 149,
         "fp": 28,
-        "fn": 36,
+        "fn": 276,
         "precision": 0.8418079096045198,
-        "recall": 0.8054054054054054,
-        "f1": 0.8232044198895028,
+        "recall": 0.35058823529411764,
+        "f1": 0.49501661129568103,
         "mean_mask_iou": 0.8955712111409039,
         "n_matched": 149
       },
       "dataset": {
-        "gold_images": 66,
+        "gold_images": 149,
         "pred_images": 120,
-        "missing_predictions": 0,
+        "missing_predictions": 83,
         "iou_threshold": 0.5
       }
     },
@@ -146,50 +146,50 @@
         "loop_all": {
           "tp": 60,
           "fp": 5,
-          "fn": 4,
+          "fn": 79,
           "precision": 0.9230769230769231,
-          "recall": 0.9375,
-          "f1": 0.9302325581395349,
+          "recall": 0.4316546762589928,
+          "f1": 0.5882352941176471,
           "mean_mask_iou": 0.8627592241834525,
           "n_matched": 60
         },
         "pin": {
           "tp": 55,
           "fp": 3,
-          "fn": 4,
+          "fn": 61,
           "precision": 0.9482758620689655,
-          "recall": 0.9322033898305084,
-          "f1": 0.94017094017094,
+          "recall": 0.47413793103448276,
+          "f1": 0.6321839080459769,
           "mean_mask_iou": 0.9206098193697871,
           "n_matched": 55
         },
         "crystal": {
           "tp": 4,
           "fp": 9,
-          "fn": 3,
+          "fn": 36,
           "precision": 0.3076923076923077,
-          "recall": 0.5714285714285714,
-          "f1": 0.4,
+          "recall": 0.1,
+          "f1": 0.15094339622641512,
           "mean_mask_iou": 0.7422849246805395,
           "n_matched": 4
         },
         "loop_face": {
           "tp": 40,
           "fp": 14,
-          "fn": 13,
+          "fn": 80,
           "precision": 0.7407407407407407,
-          "recall": 0.7547169811320755,
-          "f1": 0.7476635514018692,
+          "recall": 0.3333333333333333,
+          "f1": 0.4597701149425288,
           "mean_mask_iou": 0.8552237297527754,
           "n_matched": 40
         },
         "ice": {
           "tp": 1,
           "fp": 0,
-          "fn": 1,
+          "fn": 9,
           "precision": 1.0,
-          "recall": 0.5,
-          "f1": 0.6666666666666666,
+          "recall": 0.1,
+          "f1": 0.18181818181818182,
           "mean_mask_iou": 0.8895522388059701,
           "n_matched": 1
         }
@@ -197,17 +197,17 @@
       "overall": {
         "tp": 160,
         "fp": 31,
-        "fn": 25,
+        "fn": 265,
         "precision": 0.837696335078534,
-        "recall": 0.8648648648648649,
-        "f1": 0.851063829787234,
+        "recall": 0.3764705882352941,
+        "f1": 0.5194805194805195,
         "mean_mask_iou": 0.8779170915249036,
         "n_matched": 160
       },
       "dataset": {
-        "gold_images": 66,
+        "gold_images": 149,
         "pred_images": 120,
-        "missing_predictions": 0,
+        "missing_predictions": 83,
         "iou_threshold": 0.5
       }
     },
@@ -216,50 +216,50 @@
         "loop_all": {
           "tp": 59,
           "fp": 12,
-          "fn": 5,
+          "fn": 80,
           "precision": 0.8309859154929577,
-          "recall": 0.921875,
-          "f1": 0.874074074074074,
+          "recall": 0.4244604316546763,
+          "f1": 0.5619047619047619,
           "mean_mask_iou": 0.8836420109433821,
           "n_matched": 59
         },
         "pin": {
           "tp": 55,
           "fp": 2,
-          "fn": 4,
+          "fn": 61,
           "precision": 0.9649122807017544,
-          "recall": 0.9322033898305084,
-          "f1": 0.9482758620689654,
+          "recall": 0.47413793103448276,
+          "f1": 0.6358381502890174,
           "mean_mask_iou": 0.9211536151590175,
           "n_matched": 55
         },
         "crystal": {
           "tp": 3,
           "fp": 3,
-          "fn": 4,
+          "fn": 37,
           "precision": 0.5,
-          "recall": 0.42857142857142855,
-          "f1": 0.4615384615384615,
+          "recall": 0.075,
+          "f1": 0.13043478260869565,
           "mean_mask_iou": 0.8327909254868483,
           "n_matched": 3
         },
         "loop_face": {
           "tp": 39,
           "fp": 7,
-          "fn": 14,
+          "fn": 81,
           "precision": 0.8478260869565217,
-          "recall": 0.7358490566037735,
-          "f1": 0.7878787878787878,
+          "recall": 0.325,
+          "f1": 0.46987951807228917,
           "mean_mask_iou": 0.8596259543418816,
           "n_matched": 39
         },
         "ice": {
           "tp": 1,
           "fp": 0,
-          "fn": 1,
+          "fn": 9,
           "precision": 1.0,
-          "recall": 0.5,
-          "f1": 0.6666666666666666,
+          "recall": 0.1,
+          "f1": 0.18181818181818182,
           "mean_mask_iou": 0.8958333333333334,
           "n_matched": 1
         }
@@ -267,17 +267,17 @@
       "overall": {
         "tp": 157,
         "fp": 24,
-        "fn": 28,
+        "fn": 268,
         "precision": 0.8674033149171271,
-        "recall": 0.8486486486486486,
-        "f1": 0.8579234972677595,
+        "recall": 0.36941176470588233,
+        "f1": 0.5181518151815181,
         "mean_mask_iou": 0.8899232217103997,
         "n_matched": 157
       },
       "dataset": {
-        "gold_images": 66,
+        "gold_images": 149,
         "pred_images": 120,
-        "missing_predictions": 0,
+        "missing_predictions": 83,
         "iou_threshold": 0.5
       }
     },
@@ -286,50 +286,50 @@
         "loop_all": {
           "tp": 59,
           "fp": 8,
-          "fn": 5,
+          "fn": 80,
           "precision": 0.8805970149253731,
-          "recall": 0.921875,
-          "f1": 0.9007633587786259,
+          "recall": 0.4244604316546763,
+          "f1": 0.5728155339805825,
           "mean_mask_iou": 0.8707171878261458,
           "n_matched": 59
         },
         "pin": {
           "tp": 55,
           "fp": 4,
-          "fn": 4,
+          "fn": 61,
           "precision": 0.9322033898305084,
-          "recall": 0.9322033898305084,
-          "f1": 0.9322033898305084,
+          "recall": 0.47413793103448276,
+          "f1": 0.6285714285714284,
           "mean_mask_iou": 0.9138805674783179,
           "n_matched": 55
         },
         "crystal": {
           "tp": 4,
           "fp": 6,
-          "fn": 3,
+          "fn": 36,
           "precision": 0.4,
-          "recall": 0.5714285714285714,
-          "f1": 0.47058823529411764,
+          "recall": 0.1,
+          "f1": 0.16000000000000003,
           "mean_mask_iou": 0.7713352240501519,
           "n_matched": 4
         },
         "loop_face": {
           "tp": 40,
           "fp": 14,
-          "fn": 13,
+          "fn": 80,
           "precision": 0.7407407407407407,
-          "recall": 0.7547169811320755,
-          "f1": 0.7476635514018692,
+          "recall": 0.3333333333333333,
+          "f1": 0.4597701149425288,
           "mean_mask_iou": 0.8528345417122416,
           "n_matched": 40
         },
         "ice": {
           "tp": 1,
           "fp": 0,
-          "fn": 1,
+          "fn": 9,
           "precision": 1.0,
-          "recall": 0.5,
-          "f1": 0.6666666666666666,
+          "recall": 0.1,
+          "f1": 0.18181818181818182,
           "mean_mask_iou": 0.9156626506024096,
           "n_matched": 1
         }
@@ -337,17 +337,17 @@
       "overall": {
         "tp": 159,
         "fp": 32,
-        "fn": 26,
+        "fn": 266,
         "precision": 0.8324607329842932,
-        "recall": 0.8594594594594595,
-        "f1": 0.8457446808510638,
+        "recall": 0.37411764705882355,
+        "f1": 0.5162337662337663,
         "mean_mask_iou": 0.8789316384172502,
         "n_matched": 159
       },
       "dataset": {
-        "gold_images": 66,
+        "gold_images": 149,
         "pred_images": 120,
-        "missing_predictions": 0,
+        "missing_predictions": 83,
         "iou_threshold": 0.5
       }
     },
@@ -356,50 +356,50 @@
         "loop_all": {
           "tp": 58,
           "fp": 7,
-          "fn": 6,
+          "fn": 81,
           "precision": 0.8923076923076924,
-          "recall": 0.90625,
-          "f1": 0.8992248062015504,
+          "recall": 0.4172661870503597,
+          "f1": 0.5686274509803922,
           "mean_mask_iou": 0.873542286431727,
           "n_matched": 58
         },
         "pin": {
           "tp": 56,
           "fp": 7,
-          "fn": 3,
+          "fn": 60,
           "precision": 0.8888888888888888,
-          "recall": 0.9491525423728814,
-          "f1": 0.9180327868852458,
+          "recall": 0.4827586206896552,
+          "f1": 0.6256983240223464,
           "mean_mask_iou": 0.913231968211653,
           "n_matched": 56
         },
         "crystal": {
           "tp": 4,
           "fp": 2,
-          "fn": 3,
+          "fn": 36,
           "precision": 0.6666666666666666,
-          "recall": 0.5714285714285714,
-          "f1": 0.6153846153846153,
+          "recall": 0.1,
+          "f1": 0.1739130434782609,
           "mean_mask_iou": 0.7662505101273234,
           "n_matched": 4
         },
         "loop_face": {
           "tp": 38,
           "fp": 9,
-          "fn": 15,
+          "fn": 82,
           "precision": 0.8085106382978723,
-          "recall": 0.7169811320754716,
-          "f1": 0.76,
+          "recall": 0.31666666666666665,
+          "f1": 0.45508982035928147,
           "mean_mask_iou": 0.8637306298851367,
           "n_matched": 38
         },
         "ice": {
           "tp": 1,
           "fp": 1,
-          "fn": 1,
+          "fn": 9,
           "precision": 0.5,
-          "recall": 0.5,
-          "f1": 0.5,
+          "recall": 0.1,
+          "f1": 0.16666666666666669,
           "mean_mask_iou": 0.8808139534883721,
           "n_matched": 1
         }
@@ -407,17 +407,17 @@
       "overall": {
         "tp": 157,
         "fp": 26,
-        "fn": 28,
+        "fn": 268,
         "precision": 0.8579234972677595,
-        "recall": 0.8486486486486486,
-        "f1": 0.8532608695652173,
+        "recall": 0.36941176470588233,
+        "f1": 0.5164473684210525,
         "mean_mask_iou": 0.8826370876593987,
         "n_matched": 157
       },
       "dataset": {
-        "gold_images": 66,
+        "gold_images": 149,
         "pred_images": 120,
-        "missing_predictions": 0,
+        "missing_predictions": 83,
         "iou_threshold": 0.5
       }
     },
@@ -426,50 +426,50 @@
         "loop_all": {
           "tp": 59,
           "fp": 7,
-          "fn": 5,
+          "fn": 80,
           "precision": 0.8939393939393939,
-          "recall": 0.921875,
-          "f1": 0.9076923076923077,
+          "recall": 0.4244604316546763,
+          "f1": 0.5756097560975609,
           "mean_mask_iou": 0.8814978671212703,
           "n_matched": 59
         },
         "pin": {
           "tp": 55,
           "fp": 4,
-          "fn": 4,
+          "fn": 61,
           "precision": 0.9322033898305084,
-          "recall": 0.9322033898305084,
-          "f1": 0.9322033898305084,
+          "recall": 0.47413793103448276,
+          "f1": 0.6285714285714284,
           "mean_mask_iou": 0.9269541978233039,
           "n_matched": 55
         },
         "crystal": {
           "tp": 4,
           "fp": 4,
-          "fn": 3,
+          "fn": 36,
           "precision": 0.5,
-          "recall": 0.5714285714285714,
-          "f1": 0.5333333333333333,
+          "recall": 0.1,
+          "f1": 0.16666666666666669,
           "mean_mask_iou": 0.794146590637362,
           "n_matched": 4
         },
         "loop_face": {
           "tp": 38,
           "fp": 12,
-          "fn": 15,
+          "fn": 82,
           "precision": 0.76,
-          "recall": 0.7169811320754716,
-          "f1": 0.7378640776699028,
+          "recall": 0.31666666666666665,
+          "f1": 0.4470588235294118,
           "mean_mask_iou": 0.872694452002524,
           "n_matched": 38
         },
         "ice": {
           "tp": 1,
           "fp": 0,
-          "fn": 1,
+          "fn": 9,
           "precision": 1.0,
-          "recall": 0.5,
-          "f1": 0.6666666666666666,
+          "recall": 0.1,
+          "f1": 0.18181818181818182,
           "mean_mask_iou": 0.901840490797546,
           "n_matched": 1
         }
@@ -477,16 +477,86 @@
       "overall": {
         "tp": 157,
         "fp": 27,
-        "fn": 28,
+        "fn": 268,
         "precision": 0.8532608695652174,
-        "recall": 0.8486486486486486,
-        "f1": 0.8509485094850948,
+        "recall": 0.36941176470588233,
+        "f1": 0.51559934318555,
         "mean_mask_iou": 0.8931953571329908,
         "n_matched": 157
       },
       "dataset": {
-        "gold_images": 66,
+        "gold_images": 149,
         "pred_images": 120,
+        "missing_predictions": 83,
+        "iou_threshold": 0.5
+      }
+    },
+    "2026-09-29_v2-recipeA": {
+      "per_class": {
+        "loop_all": {
+          "tp": 134,
+          "fp": 21,
+          "fn": 5,
+          "precision": 0.864516129032258,
+          "recall": 0.9640287769784173,
+          "f1": 0.91156462585034,
+          "mean_mask_iou": 0.8689695439416028,
+          "n_matched": 134
+        },
+        "pin": {
+          "tp": 107,
+          "fp": 14,
+          "fn": 9,
+          "precision": 0.8842975206611571,
+          "recall": 0.9224137931034483,
+          "f1": 0.9029535864978903,
+          "mean_mask_iou": 0.9119533599495112,
+          "n_matched": 107
+        },
+        "crystal": {
+          "tp": 18,
+          "fp": 19,
+          "fn": 22,
+          "precision": 0.4864864864864865,
+          "recall": 0.45,
+          "f1": 0.4675324675324675,
+          "mean_mask_iou": 0.7810514836266864,
+          "n_matched": 18
+        },
+        "loop_face": {
+          "tp": 99,
+          "fp": 38,
+          "fn": 21,
+          "precision": 0.7226277372262774,
+          "recall": 0.825,
+          "f1": 0.7704280155642024,
+          "mean_mask_iou": 0.8662289384745984,
+          "n_matched": 99
+        },
+        "ice": {
+          "tp": 1,
+          "fp": 6,
+          "fn": 9,
+          "precision": 0.14285714285714285,
+          "recall": 0.1,
+          "f1": 0.11764705882352941,
+          "mean_mask_iou": 0.8793103448275862,
+          "n_matched": 1
+        }
+      },
+      "overall": {
+        "tp": 359,
+        "fp": 98,
+        "fn": 66,
+        "precision": 0.7855579868708972,
+        "recall": 0.8447058823529412,
+        "f1": 0.8140589569160998,
+        "mean_mask_iou": 0.8766457670246954,
+        "n_matched": 359
+      },
+      "dataset": {
+        "gold_images": 149,
+        "pred_images": 203,
         "missing_predictions": 0,
         "iou_threshold": 0.5
       }