Spaces:
Runtime error
Runtime error
File size: 1,393 Bytes
11120b4 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 |
# Copyright (C) 2021. Huawei Technologies Co., Ltd. All rights reserved.
# This program is free software; you can redistribute it and/or modify
# it under the terms of the MIT License.
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# MIT License for more details.
from model.utils import fix_len_compatibility
# data parameters
train_filelist_path = "../../id/train.txt"
valid_filelist_path = "../../id/valid.txt"
# test_filelist_path = 'resources/filelists/ljspeech/test.txt'
cmudict_path = "resources/ipa_dictionary_id"
add_blank = True
n_feats = 80
n_spks = 2 # 247 for Libri-TTS filelist and 1 for LJSpeech
spk_emb_dim = 64
n_feats = 80
n_fft = 1024
sample_rate = 22050
hop_length = 256
win_length = 1024
f_min = 0
f_max = 8000
# encoder parameters
n_enc_channels = 192
filter_channels = 768
filter_channels_dp = 256
n_enc_layers = 6
enc_kernel = 3
enc_dropout = 0.1
n_heads = 2
window_size = 4
# decoder parameters
dec_dim = 64
beta_min = 0.05
beta_max = 20.0
pe_scale = 1000 # 1 for `grad-tts-old.pt` checkpoint
# training parameters
log_dir = "logs/grad-tts-bookbot-id"
test_size = 4
n_epochs = 2000
batch_size = 16
learning_rate = 1e-4
seed = 37
save_every = 100
out_size = fix_len_compatibility(2 * 22050 // 256)
num_workers = 6
|