eval_metrics.json

{
  "n": 500,
  "threshold": 0.5,
  "max_seq_length": 3072,
  "is_valid_accuracy": 0.958,
  "category_set_accuracy": 0.82,
  "micro_f1": 0.8976234003656307,
  "macro_f1": 0.8954906015550027,
  "per_language_f1": {
    "Python": 0.8627450980392157,
    "JavaScript": 0.8163265306122449,
    "Java": 0.9166666666666666,
    "C": 0.8636363636363636,
    "C++": 0.9361702127659575,
    "C#": 0.926829268292683,
    "Go": 0.918918918918919,
    "Rust": 0.9811320754716981,
    "Kotlin": 1.0,
    "Swift": 0.9166666666666666,
    "Ruby": 0.9,
    "R": 0.9056603773584906,
    "Scala": 0.7619047619047619,
    "Perl": 0.8571428571428571,
    "Lua": 0.8666666666666667,
    "Bash": 0.7222222222222222,
    "PowerShell": 0.8333333333333334,
    "Batch": 0.9019607843137255,
    "SQL": 0.9803921568627451,
    "Dockerfile": 0.9767441860465116,
    "YAML": 0.9545454545454546,
    "Makefile": 0.8780487804878049,
    "Terraform": 0.8947368421052632,
    "AWK": 0.9259259259259259,
    "jq": 0.8888888888888888
  },
  "latency_ms_per_example": {
    "mean": 2.3932456970214844,
    "p95": 3.833106905221939,
    "device": "cuda:0"
  },
  "base_model": "jhu-clsp/mmBERT-base",
  "epochs": 2
}