vocoder/config.json

{
  "architectures": [
    "CodeHiFiGANModel"
  ],
  "duration_predictor": true,
  "duration_predictor_activation": "relu",
  "embedding_dim": 128,
  "encoder_embed_dim": 128,
  "f0": false,
  "f0_quant_num_bin": 0,
  "initializer_range": 0.02,
  "leaky_relu_slope": 0.1,
  "model_in_dim": 128,
  "multispeaker": false,
  "num_embeddings": 1000,
  "num_speakers": 200,
  "resblock_dilation_sizes": [
    [
      1,
      3,
      5
    ],
    [
      1,
      3,
      5
    ],
    [
      1,
      3,
      5
    ]
  ],
  "resblock_kernel_sizes": [
    3,
    7,
    11
  ],
  "sampling_rate": 16000,
  "speaker_embedding": false,
  "speaker_embedding_dim": 256,
  "torch_dtype": "float32",
  "transformers_version": "4.22.0.dev0",
  "upsample_initial_channel": 512,
  "upsample_kernel_sizes": [
    11,
    8,
    8,
    4,
    4
  ],
  "upsample_rates": [
    5,
    4,
    4,
    2,
    2
  ],
  "variance_predictor_dropout": 0.5,
  "variance_predictor_hidden_dim": 128,
  "variance_predictor_kernel_size": 3
}