configs/vits_pre.json

{
  "data": {
    "add_blank": true,
    "cleaned_text": true,
    "filter_length": 1024,
    "hop_length": 256,
    "max_wav_value": 32768.0,
    "mel_fmax": null,
    "mel_fmin": 0.0,
    "n_mel_channels": 80,
    "n_speakers": 8,
    "sampling_rate": 22050,
    "text_cleaners": [
      "zh_ja_mixture_cleaners"
    ],
    "training_files": "/root/content/vits/filelists/vits_pre_train.txt.cleaned",
    "validation_files": "/root/content/vits/filelists/vits_pre_val.txt.cleaned",
    "win_length": 1024
  },
  "model": {
    "filter_channels": 768,
    "gin_channels": 256,
    "hidden_channels": 192,
    "inter_channels": 192,
    "kernel_size": 3,
    "n_heads": 2,
    "n_layers": 6,
    "n_layers_q": 3,
    "p_dropout": 0.1,
    "resblock": "1",
    "resblock_dilation_sizes": [
      [
        1,
        3,
        5
      ],
      [
        1,
        3,
        5
      ],
      [
        1,
        3,
        5
      ]
    ],
    "resblock_kernel_sizes": [
      3,
      7,
      11
    ],
    "upsample_initial_channel": 512,
    "upsample_kernel_sizes": [
      16,
      16,
      4,
      4
    ],
    "upsample_rates": [
      8,
      8,
      2,
      2
    ],
    "use_spectral_norm": false
  },
  "speakers": [
    "yilanqiu",
    "xing",
    "speechocean",
    "jsut"
  ],
  "symbols": [
    "_",
    ",",
    ".",
    "!",
    "?",
    "-",
    "~",
    "\u2026",
    "A",
    "E",
    "I",
    "N",
    "O",
    "Q",
    "U",
    "a",
    "b",
    "d",
    "e",
    "f",
    "g",
    "h",
    "i",
    "j",
    "k",
    "l",
    "m",
    "n",
    "o",
    "p",
    "r",
    "s",
    "t",
    "u",
    "v",
    "w",
    "y",
    "z",
    "\u0283",
    "\u02a7",
    "\u02a6",
    "\u026f",
    "\u0279",
    "\u0259",
    "\u0265",
    "\u207c",
    "\u02b0",
    "`",
    "\u2192",
    "\u2193",
    "\u2191",
    " "
  ],
  "train": {
    "batch_size": 32,
    "betas": [
      0.8,
      0.99
    ],
    "c_kl": 1.0,
    "c_mel": 45,
    "epochs": 5000,
    "eps": 1e-09,
    "eval_interval": 1000,
    "fp16_run": true,
    "init_lr_ratio": 1,
    "learning_rate": 0.0002,
    "log_interval": 200,
    "lr_decay": 0.999875,
    "seed": 1234,
    "segment_size": 8192,
    "warmup_epochs": 0
  }
}