{
  "assamese": {
    "n_samples": 500,
    "wer_raw": 32.73,
    "wer_norm": 25.11,
    "wer_numcanon": 25.11,
    "mer": 11.64,
    "cer_norm": 10.73,
    "empty_hypotheses": 22,
    "normalization_delta": {
      "raw_to_norm": 7.61,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 13.48
    },
    "wer_norm_nonum": 25.08,
    "mer_nonum": 11.37,
    "numeric_samples_dropped": 2,
    "space_norm_wer": 21.56
  },
  "bengali": {
    "n_samples": 500,
    "wer_raw": 11.19,
    "wer_norm": 10.44,
    "wer_numcanon": 10.44,
    "mer": 2.08,
    "cer_norm": 2.17,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.74,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 8.37
    },
    "wer_norm_nonum": 10.44,
    "mer_nonum": 2.08,
    "numeric_samples_dropped": 0,
    "space_norm_wer": 6.44
  },
  "english": {
    "n_samples": 500,
    "wer_raw": 24.83,
    "wer_norm": 13.94,
    "wer_numcanon": 13.94,
    "mer": 9.33,
    "cer_norm": 9.28,
    "empty_hypotheses": 31,
    "normalization_delta": {
      "raw_to_norm": 10.9,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 4.61
    },
    "wer_norm_nonum": 10.23,
    "mer_nonum": 5.54,
    "numeric_samples_dropped": 97,
    "space_norm_wer": 9.11
  },
  "gujarati": {
    "n_samples": 500,
    "wer_raw": 17.28,
    "wer_norm": 13.1,
    "wer_numcanon": 13.1,
    "mer": 4.44,
    "cer_norm": 4.13,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 4.18,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 8.66
    },
    "wer_norm_nonum": 13.1,
    "mer_nonum": 4.44,
    "numeric_samples_dropped": 0,
    "space_norm_wer": 9.97
  },
  "hindi": {
    "n_samples": 500,
    "wer_raw": 15.56,
    "wer_norm": 9.33,
    "wer_numcanon": 9.33,
    "mer": 5.9,
    "cer_norm": 5.1,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 6.23,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 3.43
    },
    "wer_norm_nonum": 9.33,
    "mer_nonum": 5.9,
    "numeric_samples_dropped": 0,
    "space_norm_wer": 8.17
  },
  "kannada": {
    "n_samples": 500,
    "wer_raw": 23.15,
    "wer_norm": 18.56,
    "wer_numcanon": 18.56,
    "mer": 3.49,
    "cer_norm": 3.71,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 4.59,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 15.07
    },
    "wer_norm_nonum": 18.32,
    "mer_nonum": 3.26,
    "numeric_samples_dropped": 1,
    "space_norm_wer": 11.73
  },
  "malayalam": {
    "n_samples": 500,
    "wer_raw": 34.17,
    "wer_norm": 31.68,
    "wer_numcanon": 31.68,
    "mer": 5.9,
    "cer_norm": 5.9,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 2.48,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 25.78
    },
    "wer_norm_nonum": 31.68,
    "mer_nonum": 5.9,
    "numeric_samples_dropped": 0,
    "space_norm_wer": 24.44
  },
  "marathi": {
    "n_samples": 500,
    "wer_raw": 18.78,
    "wer_norm": 14.83,
    "wer_numcanon": 14.83,
    "mer": 3.65,
    "cer_norm": 3.68,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 3.96,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 11.18
    },
    "wer_norm_nonum": 14.83,
    "mer_nonum": 3.65,
    "numeric_samples_dropped": 0,
    "space_norm_wer": 10.24
  },
  "odia": {
    "n_samples": 500,
    "wer_raw": 17.44,
    "wer_norm": 17.03,
    "wer_numcanon": 17.03,
    "mer": 4.28,
    "cer_norm": 4.35,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.41,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 12.75
    },
    "wer_norm_nonum": 17.03,
    "mer_nonum": 4.28,
    "numeric_samples_dropped": 0,
    "space_norm_wer": 10.78
  },
  "punjabi": {
    "n_samples": 500,
    "wer_raw": 17.32,
    "wer_norm": 12.35,
    "wer_numcanon": 12.35,
    "mer": 4.24,
    "cer_norm": 3.87,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 4.97,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 8.11
    },
    "wer_norm_nonum": 12.35,
    "mer_nonum": 4.24,
    "numeric_samples_dropped": 0,
    "space_norm_wer": 9.87
  },
  "tamil": {
    "n_samples": 500,
    "wer_raw": 23.08,
    "wer_norm": 21.9,
    "wer_numcanon": 21.9,
    "mer": 4.15,
    "cer_norm": 4.26,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 1.19,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 17.75
    },
    "wer_norm_nonum": 21.9,
    "mer_nonum": 4.15,
    "numeric_samples_dropped": 0,
    "space_norm_wer": 14.31
  },
  "telugu": {
    "n_samples": 500,
    "wer_raw": 25.07,
    "wer_norm": 23.29,
    "wer_numcanon": 23.29,
    "mer": 5.71,
    "cer_norm": 5.9,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 1.78,
      "norm_to_numcanon": 0.0,
      "norm_to_mer": 17.58
    },
    "wer_norm_nonum": 23.29,
    "mer_nonum": 5.71,
    "numeric_samples_dropped": 0,
    "space_norm_wer": 13.77
  },
  "__overall__": {
    "n_samples": 6000,
    "wer_raw": 21.94,
    "wer_norm": 17.74,
    "wer_numcanon": 17.74,
    "mer": 5.36,
    "cer_norm": 5.28,
    "wer_norm_nonum": 17.55,
    "mer_nonum": 5.07,
    "numeric_samples_dropped": 100,
    "space_norm_wer": 12.73
  },
  "__macro_avg__": {
    "n_languages": 12,
    "wer_raw": 21.72,
    "wer_norm": 17.63,
    "wer_numcanon": 17.63,
    "mer": 5.4,
    "cer_norm": 5.26,
    "wer_norm_nonum": 17.3,
    "mer_nonum": 5.04,
    "space_norm_wer": 12.53
  },
  "__meta__": {
    "checkpoint": "api/elevenlabs-scribe-v2",
    "checkpoint_name": "baseline",
    "model_id": "elevenlabs-scribe-v2",
    "model_type": "elevenlabs-scribe-v2",
    "dataset": "BayAreaBoys/indic-asr-benchmark-6k",
    "batch_size": 1,
    "inference_time_sec": 0,
    "total_audio_sec": 40354.42,
    "rtf": 0,
    "timestamp": "2026-03-27T07:17:28Z",
    "normalization_version": "v1",
    "framework": "api"
  }
}