pits / configs /config_en.yaml
junhyouk lee
hfdemo
b8b70ac
train:
log_interval: 200 # step unit
eval_interval: 400 # step unit
save_interval: 50 # epoch unit: 50 for baseline / 500 for fine-tuning
seed: 1234
epochs: 7000
learning_rate: 2e-4
betas: [0.8, 0.99]
eps: 1e-9
batch_size: 48
fp16_run: True #False
lr_decay: 0.999875
segment_size: 8192
c_mel: 45
c_kl: 1.0
c_vq: 1.
c_commit: 0.2
c_yin: 45.
log_path: "/pits/logs"
n_sample: 3
alpha: 200
data:
data_path: "/DATA/audio/VCTK-0.92"
training_files: "filelists/vctk_train_g2p.txt"
validation_files: "filelists/vctk_val_g2p.txt"
languages: "en_US"
text_cleaners: ["english_cleaners"]
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
speakers: ["p225", "p226", "p227", "p228", "p229", "p230", "p231", "p232", "p233", "p234", "p236", "p237", "p238", "p239", "p240", "p241", "p243", "p244", "p245", "p246", "p247", "p248", "p249", "p250", "p251", "p252", "p253", "p254", "p255", "p256", "p257", "p258", "p259", "p260", "p261", "p262", "p263", "p264", "p265", "p266", "p267", "p268", "p269", "p270", "p271", "p272", "p273", "p274", "p275", "p276", "p277", "p278", "p279", "p281", "p282", "p283", "p284", "p285", "p286", "p287", "p288", "p292", "p293", "p294", "p295", "p297", "p298", "p299", "p300", "p301", "p302", "p303", "p304", "p305", "p306", "p307", "p308", "p310", "p311", "p312", "p313", "p314", "p316", "p317", "p318", "p323", "p326", "p329", "p330", "p333", "p334", "p335", "p336", "p339", "p340", "p341", "p343", "p345", "p347", "p351", "p360", "p361", "p362", "p363", "p364", "p374", "p376", "s5"]
persistent_workers: True
midi_start: -5
midi_end: 75
midis: 80
ying_window: 2048
ying_hop: 256
tau_max: 2048
octave_range: 24
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
gin_channels: 256
codebook_size: 320
yin_channels: 80
yin_start: 15 # scope start bin in nansy = 1.5/8
yin_scope: 50 # scope ratio in nansy = 5/8
yin_shift_range: 15 # same as default start index of yingram