config.json

{
  "architectures": [
    "BarkModel"
  ],
  "coarse_acoustics_config": {
    "architectures": [
      "BarkCoarseModel"
    ],
    "bias": false,
    "block_size": 1024,
    "dropout": 0.0,
    "dtype": "float32",
    "hidden_size": 768,
    "initializer_range": 0.02,
    "input_vocab_size": 12096,
    "model_type": "coarse_acoustics",
    "num_heads": 12,
    "num_layers": 12,
    "output_vocab_size": 12096,
    "use_cache": true
  },
  "codec_config": {
    "_name_or_path": "facebook/encodec_24khz",
    "architectures": [
      "EncodecModel"
    ],
    "audio_channels": 1,
    "chunk_length_s": null,
    "codebook_dim": 128,
    "codebook_size": 1024,
    "compress": 2,
    "dilation_growth_rate": 2,
    "dtype": "float32",
    "hidden_size": 128,
    "kernel_size": 7,
    "last_kernel_size": 7,
    "model_type": "encodec",
    "norm_type": "weight_norm",
    "normalize": false,
    "num_filters": 32,
    "num_lstm_layers": 2,
    "num_residual_layers": 1,
    "overlap": null,
    "pad_mode": "reflect",
    "residual_kernel_size": 3,
    "sampling_rate": 24000,
    "target_bandwidths": [
      1.5,
      3.0,
      6.0,
      12.0,
      24.0
    ],
    "trim_right_ratio": 1.0,
    "upsampling_ratios": [
      8,
      5,
      4,
      2
    ],
    "use_causal_conv": true,
    "use_conv_shortcut": true
  },
  "dtype": "float32",
  "fine_acoustics_config": {
    "architectures": [
      "BarkFineModel"
    ],
    "bias": false,
    "block_size": 1024,
    "dropout": 0.0,
    "dtype": "float32",
    "hidden_size": 768,
    "initializer_range": 0.02,
    "input_vocab_size": 1056,
    "model_type": "fine_acoustics",
    "n_codes_given": 1,
    "n_codes_total": 8,
    "num_heads": 12,
    "num_layers": 12,
    "output_vocab_size": 1056,
    "use_cache": true
  },
  "initializer_range": 0.02,
  "model_type": "bark",
  "semantic_config": {
    "architectures": [
      "BarkSemanticModel"
    ],
    "bias": false,
    "block_size": 1024,
    "dropout": 0.0,
    "dtype": "float32",
    "hidden_size": 768,
    "initializer_range": 0.02,
    "input_vocab_size": 129600,
    "model_type": "semantic",
    "num_heads": 12,
    "num_layers": 12,
    "output_vocab_size": 10048,
    "use_cache": true
  },
  "transformers_version": "4.57.0"
}