mirror of
https://github.com/w-okada/voice-changer.git
synced 2025-02-03 08:43:57 +03:00
111 lines
2.3 KiB
JSON
111 lines
2.3 KiB
JSON
|
{
|
||
|
|
"train": {
|
||
|
|
"log_interval": 1000,
|
||
|
|
"eval_interval": 2000,
|
||
|
|
"seed": 1234,
|
||
|
|
"epochs": 10000,
|
||
|
|
"learning_rate": 0.0002,
|
||
|
|
"betas": [
|
||
|
|
0.8,
|
||
|
|
0.99
|
||
|
|
],
|
||
|
|
"eps": 1e-09,
|
||
|
|
"batch_size": 10,
|
||
|
|
"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": "filelists/dec_not_propagation_label_and_change_melspec_textful.txt",
|
||
|
|
"validation_files": "filelists/dec_not_propagation_label_and_change_melspec_textful_val.txt",
|
||
|
|
"training_files_notext": "filelists/dec_not_propagation_label_and_change_melspec_textless.txt",
|
||
|
|
"validation_files_notext": "filelists/dec_not_propagation_label_and_change_melspec_val_textless.txt",
|
||
|
|
"text_cleaners": [
|
||
|
|
"japanese_cleaners"
|
||
|
|
],
|
||
|
|
"max_wav_value": 32768.0,
|
||
|
|
"sampling_rate": 24000,
|
||
|
|
"filter_length": 512,
|
||
|
|
"hop_length": 128,
|
||
|
|
"win_length": 512,
|
||
|
|
"n_mel_channels": 80,
|
||
|
|
"mel_fmin": 0.0,
|
||
|
|
"mel_fmax": null,
|
||
|
|
"add_blank": true,
|
||
|
|
"n_speakers": 110,
|
||
|
|
"cleaned_text": false
|
||
|
|
},
|
||
|
|
"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,
|
||
|
|
4,
|
||
|
|
2,
|
||
|
|
2
|
||
|
|
],
|
||
|
|
"upsample_initial_channel": 512,
|
||
|
|
"upsample_kernel_sizes": [
|
||
|
|
16,
|
||
|
|
16,
|
||
|
|
8,
|
||
|
|
8
|
||
|
|
],
|
||
|
|
"n_layers_q": 3,
|
||
|
|
"use_spectral_norm": false,
|
||
|
|
"n_flow": 8,
|
||
|
|
"gin_channels": 256
|
||
|
|
},
|
||
|
|
"others": {
|
||
|
|
"os_type": "linux"
|
||
|
|
},
|
||
|
|
"augmentation": {
|
||
|
|
"enable" : true,
|
||
|
|
"gain_p" : 0.5,
|
||
|
|
"min_gain_in_db" : -10,
|
||
|
|
"max_gain_in_db" : 10,
|
||
|
|
"time_stretch_p" : 0.5,
|
||
|
|
"min_rate" : 0.75,
|
||
|
|
"max_rate" : 1.25,
|
||
|
|
"pitch_shift_p" : 0.0,
|
||
|
|
"min_semitones" : -4.0,
|
||
|
|
"max_semitones" : 4.0,
|
||
|
|
"add_gaussian_noise_p" : 0.0,
|
||
|
|
"min_amplitude" : 0.001,
|
||
|
|
"max_amplitude" : 0.04,
|
||
|
|
"frequency_mask_p" : 0.0
|
||
|
|
}
|
||
|
|
}
|