{
  "config": {
    "model": "google/gemma-4-12B-it",
    "dataset": "earnings22",
    "split": "test",
    "backend": "vllm",
    "quantization": null,
    "limit": null,
    "seed": 42,
    "output_dir": "eval_results",
    "output_json": null,
    "output_csv": null,
    "quiet": true,
    "streaming": false,
    "prompt": "Transcribe the following speech segment in its original language.",
    "max_output_tokens": null,
    "chunk_duration_s": null,
    "batch_size": 8
  },
  "corpus_metrics": {
    "wer": 0.5039571361377533,
    "cer": 0.3854952347820768,
    "mer": 0.46497103718938093,
    "wil": 0.6161775635312747,
    "substitution_rate": 0.21237653121740732,
    "insertion_rate": 0.08384629542526432,
    "deletion_rate": 0.2077343094950817,
    "latency": {
      "mean": 0.1244895535550345,
      "p50": 0.10963858262402937,
      "p95": 0.22648817962181056
    },
    "rtfx": {
      "mean": 63.01479438472583,
      "p50": 53.811823156995786,
      "p95": 149.9791398648187
    },
    "audio_duration": {
      "mean": 7.12854683509668,
      "min": 0.2200625,
      "max": 36.36,
      "p50": 6.14,
      "total": 19539.346875
    },
    "num_samples": 2741,
    "bleu": null,
    "throughput_rtfx": 54.88587640387496
  },
  "sample_results": []
}