aggregate_metrics.json

{
  "dataset": "ViFN",
  "dataset_rows": 1406,
  "split_counts": {
    "train": 1124,
    "validation": 141,
    "test": 141
  },
  "split_seed": 42,
  "source_sha256": "80ca7eab9383bb0487e64c857b81956f3abd5147aaa19888d4ca8cdff7fd3a50",
  "canonical_data_sha256": "dcf674944d5a4c323f1817ff33c088ff41e427b26b9892e31cae70c077f6e5f0",
  "base_model": "vinai/phobert-base",
  "base_revision": "c1e37c5c86f918761049cef6fa216b4779d0d01d",
  "fine_tuning_seeds": [
    42,
    22,
    202
  ],
  "representative_seed": 22,
  "selection_rule": "highest development Macro-F1; test metrics are not used for selection",
  "aggregate_metrics": {
    "dev_accuracy": {
      "mean": 0.8770685579196217,
      "std": 0.008189365520420171,
      "values": [
        0.8865248226950354,
        0.8723404255319149,
        0.8723404255319149
      ]
    },
    "dev_macro_precision": {
      "mean": 0.878471904660285,
      "std": 0.008271926112746478,
      "values": [
        0.8878787878787879,
        0.8723340040241448,
        0.875202922077922
      ]
    },
    "dev_macro_recall": {
      "mean": 0.8768947015425889,
      "std": 0.008162130821931303,
      "values": [
        0.886317907444668,
        0.8723340040241448,
        0.8720321931589538
      ]
    },
    "dev_macro_f1": {
      "mean": 0.8769136547163822,
      "std": 0.008201239636128432,
      "values": [
        0.8863819500402901,
        0.8723340040241448,
        0.8720250100847116
      ]
    },
    "dev_weighted_f1": {
      "mean": 0.876940339845822,
      "std": 0.008202534495814187,
      "values": [
        0.8864105245712393,
        0.8723404255319149,
        0.872070069434312
      ]
    },
    "test_accuracy": {
      "mean": 0.8605200945626477,
      "std": 0.004094682760210118,
      "values": [
        0.8652482269503546,
        0.8581560283687943,
        0.8581560283687943
      ]
    },
    "test_macro_precision": {
      "mean": 0.8635432700730278,
      "std": 0.00454023255846262,
      "values": [
        0.866969696969697,
        0.8583937198067633,
        0.865266393442623
      ]
    },
    "test_macro_recall": {
      "mean": 0.8607981220657277,
      "std": 0.004070822509857444,
      "values": [
        0.8654929577464789,
        0.8582494969818913,
        0.8586519114688129
      ]
    },
    "test_macro_f1": {
      "mean": 0.8602881146210591,
      "std": 0.004211351351138073,
      "values": [
        0.865139692927259,
        0.8581488933601609,
        0.8575757575757577
      ]
    },
    "test_weighted_f1": {
      "mean": 0.8602552002720701,
      "std": 0.004218391661333309,
      "values": [
        0.865112559421485,
        0.8581417583515276,
        0.8575112830431979
      ]
    }
  },
  "training": {
    "epochs": 3,
    "learning_rate": 2e-05,
    "weight_decay": 0.01,
    "warmup_ratio": 0.1,
    "effective_batch_size": 8,
    "max_length": 256,
    "text_mode": "segmented",
    "class_weighting": false,
    "resampling": false
  },
  "environment": {
    "python": "3.12.13",
    "execution_mode": "full"
  },
  "created_at_utc": "2026-07-18T17:59:52.912002+00:00"
}