{
  "assamese": {
    "n_samples": 500,
    "wer_raw": 33.17,
    "wer_norm": 32.88,
    "wer_numcanon": 32.88,
    "space_norm_wer": 28.47,
    "mer": 14.99,
    "cer_norm": 13.97,
    "empty_hypotheses": 1,
    "normalization_delta": {
      "raw_to_norm": 0.3,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 4.41,
      "norm_to_mer": 17.88
    }
  },
  "bengali": {
    "n_samples": 500,
    "wer_raw": 32.35,
    "wer_norm": 31.59,
    "wer_numcanon": 31.59,
    "space_norm_wer": 23.63,
    "mer": 8.75,
    "cer_norm": 8.75,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.77,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 7.96,
      "norm_to_mer": 22.83
    }
  },
  "english": {
    "n_samples": 500,
    "wer_raw": 30.74,
    "wer_norm": 18.22,
    "wer_numcanon": 18.22,
    "space_norm_wer": 13.53,
    "mer": 13.35,
    "cer_norm": 12.66,
    "empty_hypotheses": 21,
    "normalization_delta": {
      "raw_to_norm": 12.52,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 4.7,
      "norm_to_mer": 4.87
    }
  },
  "gujarati": {
    "n_samples": 500,
    "wer_raw": 25.88,
    "wer_norm": 25.76,
    "wer_numcanon": 25.76,
    "space_norm_wer": 19.68,
    "mer": 7.4,
    "cer_norm": 7.32,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.12,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 6.08,
      "norm_to_mer": 18.36
    }
  },
  "hindi": {
    "n_samples": 500,
    "wer_raw": 18.43,
    "wer_norm": 18.38,
    "wer_numcanon": 18.38,
    "space_norm_wer": 13.85,
    "mer": 6.56,
    "cer_norm": 6.25,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.05,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 4.53,
      "norm_to_mer": 11.82
    }
  },
  "kannada": {
    "n_samples": 500,
    "wer_raw": 42.29,
    "wer_norm": 42.04,
    "wer_numcanon": 42.04,
    "space_norm_wer": 31.56,
    "mer": 9.51,
    "cer_norm": 9.75,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.25,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 10.48,
      "norm_to_mer": 32.53
    }
  },
  "malayalam": {
    "n_samples": 500,
    "wer_raw": 48.86,
    "wer_norm": 48.03,
    "wer_numcanon": 48.03,
    "space_norm_wer": 37.22,
    "mer": 9.54,
    "cer_norm": 9.6,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.83,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 10.81,
      "norm_to_mer": 38.49
    }
  },
  "marathi": {
    "n_samples": 500,
    "wer_raw": 39.41,
    "wer_norm": 39.24,
    "wer_numcanon": 39.24,
    "space_norm_wer": 29.65,
    "mer": 11.34,
    "cer_norm": 11.47,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.18,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 9.58,
      "norm_to_mer": 27.9
    }
  },
  "odia": {
    "n_samples": 500,
    "wer_raw": 34.84,
    "wer_norm": 34.27,
    "wer_numcanon": 34.27,
    "space_norm_wer": 24.41,
    "mer": 9.14,
    "cer_norm": 9.24,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.57,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 9.85,
      "norm_to_mer": 25.13
    }
  },
  "punjabi": {
    "n_samples": 500,
    "wer_raw": 24.42,
    "wer_norm": 24.39,
    "wer_numcanon": 24.39,
    "space_norm_wer": 18.43,
    "mer": 7.39,
    "cer_norm": 7.16,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.03,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 5.96,
      "norm_to_mer": 17.0
    }
  },
  "tamil": {
    "n_samples": 500,
    "wer_raw": 44.09,
    "wer_norm": 43.66,
    "wer_numcanon": 43.66,
    "space_norm_wer": 33.29,
    "mer": 9.74,
    "cer_norm": 9.8,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 0.43,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 10.37,
      "norm_to_mer": 33.92
    }
  },
  "telugu": {
    "n_samples": 500,
    "wer_raw": 34.23,
    "wer_norm": 32.82,
    "wer_numcanon": 32.82,
    "space_norm_wer": 22.55,
    "mer": 6.46,
    "cer_norm": 6.76,
    "empty_hypotheses": 0,
    "normalization_delta": {
      "raw_to_norm": 1.41,
      "norm_to_numcanon": 0.0,
      "norm_to_space_norm": 10.27,
      "norm_to_mer": 26.36
    }
  },
  "__overall__": {
    "n_samples": 6000,
    "wer_raw": 33.89,
    "wer_norm": 32.5,
    "wer_numcanon": 32.5,
    "space_norm_wer": 24.71,
    "mer": 9.59,
    "cer_norm": 9.51
  },
  "__macro_avg__": {
    "n_languages": 12,
    "wer_raw": 34.06,
    "wer_norm": 32.61,
    "wer_numcanon": 32.61,
    "space_norm_wer": 24.69,
    "mer": 9.51,
    "cer_norm": 9.39
  },
  "__meta__": {
    "checkpoint": "/home/ubuntu/training/checkpoints/maya-asr-tdt-1.1b-ckpt-260000/model.ckpt",
    "checkpoint_name": "ckpt-260000",
    "model_id": "parakeet-tdt-1.1b-lang",
    "model_type": "maya-asr-hybrid-tdt-1.1b",
    "dataset": "BayAreaBoys/indic-asr-benchmark-6k",
    "batch_size": 32,
    "inference_time_sec": 360.98,
    "total_audio_sec": 40354.46,
    "rtf": 0.0089,
    "timestamp": "2026-03-30T07:13:36Z",
    "gpu": "NVIDIA A100-SXM4-80GB",
    "framework": "nemo",
    "normalization_version": "v1",
    "jiwer_version": "unknown"
  }
}