model_config.json

{
  "model": {
    "n_mels": 80,
    "feature_normalize": "utterance",
    "d_model": 384,
    "n_layers": 12,
    "n_heads": 6,
    "ff_multiplier": 3.0,
    "conv_kernel": 31,
    "conv_expansion": 2,
    "conv_norm": "batch",
    "se_ratio": 4,
    "dropout": 0.1,
    "stochastic_depth": 0.1,
    "vocab_size": 181,
    "blank_bias_init": 0.0,
    "aux_ctc_layers": [
      6
    ],
    "self_conditioning": false,
    "languages": [
      "urd",
      "snd"
    ],
    "language_embedding": true,
    "lid_head": true,
    "gradient_checkpointing": false
  },
  "parameters": {
    "subsampling": 4281216,
    "blocks": 53455104,
    "ctc_head": 69685,
    "aux_heads": 69685,
    "self_conditioning": 0,
    "language_embedding": 1152,
    "lid_head": 1155,
    "total": 57877997
  }
}