evaluation/evaluation_metrics.json

{
  "overall": {
    "accuracy": 0.06,
    "precision_macro": 0.25,
    "recall_macro": 0.42,
    "f1_macro": 0.28
  },
  "per_class": {
    "precision": {
      "games": 0.38,
      "geometry": 0.1,
      "graphs": 0.16,
      "math": 0.26,
      "number theory": 0.32,
      "other": 0.0,
      "probabilities": 0.24,
      "strings": 0.44,
      "trees": 0.36
    },
    "recall": {
      "games": 0.13,
      "geometry": 0.62,
      "graphs": 0.51,
      "math": 0.68,
      "number theory": 0.42,
      "other": 0.0,
      "probabilities": 0.39,
      "strings": 0.47,
      "trees": 0.51
    },
    "f1": {
      "games": 0.19,
      "geometry": 0.17,
      "graphs": 0.25,
      "math": 0.38,
      "number theory": 0.36,
      "other": 0.0,
      "probabilities": 0.3,
      "strings": 0.46,
      "trees": 0.43
    }
  }
}