VITS-yasmin / config.json
huseinzol05's picture
Upload config.json
dccff6d
raw
history blame
1.38 kB
{
"train": {
"log_interval": 2,
"eval_interval": 2500,
"seed": 1234,
"epochs": 20000,
"learning_rate": 2e-4,
"betas": [0.8, 0.99],
"eps": 1e-9,
"batch_size": 32,
"fp16_run": true,
"lr_decay": 0.999875,
"segment_size": 8192,
"init_lr_ratio": 1,
"warmup_epochs": 0,
"c_mel": 45,
"c_kl": 1.0
},
"data": {
"training_files":"/home/husein/speech-bahasa/yasmin-vits.json",
"validation_files":"/home/husein/speech-bahasa/yasmin-vits.json",
"text_cleaners":[""],
"max_wav_value": 32768.0,
"sampling_rate": 22050,
"filter_length": 1024,
"hop_length": 256,
"win_length": 1024,
"n_mel_channels": 80,
"mel_fmin": 0.0,
"mel_fmax": null,
"add_blank": true,
"n_speakers": 0,
"cleaned_text": true
},
"model": {
"inter_channels": 192,
"hidden_channels": 192,
"filter_channels": 768,
"n_heads": 2,
"n_layers": 6,
"kernel_size": 3,
"p_dropout": 0.1,
"resblock": "1",
"resblock_kernel_sizes": [3,7,11],
"resblock_dilation_sizes": [[1,3,5], [1,3,5], [1,3,5]],
"upsample_rates": [8,8,2,2],
"upsample_initial_channel": 512,
"upsample_kernel_sizes": [16,16,4,4],
"n_layers_q": 3,
"use_spectral_norm": false
}
}