config.json

{
    "process_pids": [
        38914,
        39245,
        39498
    ],
    "train": {
        "log_interval": 10,
        "seed": 6767,
        "learning_rate": 1e-05,
        "betas": [
            0.8,
            0.99
        ],
        "eps": 1e-09,
        "lr_decay": 0.9995,
        "segment_size": 17280,
        "c_mel": 35,
        "c_kl": 1.0,
        "grad_clip": 1.0,
        "use_kahan_summation": false,
        "momentum_dtype": "torch.float32",
        "variance_dtype": "torch.float32",
        "compensation_buffer_dtype": "torch.float32"
    },
    "data": {
        "max_wav_value": 32768.0,
        "sample_rate": 48000,
        "filter_length": 2048,
        "hop_length": 480,
        "win_length": 2048,
        "n_mel_channels": 128,
        "mel_fmin": 0.0,
        "mel_fmax": 24000
    },
    "model": {
        "inter_channels": 512,
        "hidden_channels": 512,
        "filter_channels": 2048,
        "text_enc_hidden_dim": 768,
        "n_heads": 4,
        "n_layers": 8,
        "kernel_size": 3,
        "p_dropout": 0,
        "resblock": "1",
        "resblock_kernel_sizes": [
            3,
            7,
            11,
            15
        ],
        "resblock_dilation_sizes": [
            [
                1,
                3,
                5,
                9
            ],
            [
                1,
                3,
                5,
                9
            ],
            [
                1,
                3,
                5,
                9
            ]
        ],
        "upsample_rates": [
            12,
            10,
            2,
            2
        ],
        "upsample_initial_channel": 1024,
        "upsample_kernel_sizes": [
            24,
            20,
            4,
            4
        ],
        "use_spectral_norm": true,
        "gin_channels": 256,
        "spk_embed_dim": 109
    }
}