|
|
|
dataset_conf: |
|
|
|
batch_size: 64 |
|
|
|
num_speakers: 3242 |
|
|
|
num_workers: 4 |
|
|
|
min_duration: 0.5 |
|
|
|
max_duration: 3 |
|
|
|
do_vad: False |
|
|
|
sample_rate: 16000 |
|
|
|
use_dB_normalization: True |
|
|
|
target_dB: -20 |
|
|
|
train_list: 'dataset/train_list.txt' |
|
|
|
test_list: 'dataset/test_list.txt' |
|
|
|
|
|
preprocess_conf: |
|
|
|
feature_method: 'MelSpectrogram' |
|
|
|
|
|
feature_conf: |
|
sample_rate: 16000 |
|
n_fft: 1024 |
|
hop_length: 320 |
|
win_length: 1024 |
|
f_min: 50.0 |
|
f_max: 14000.0 |
|
n_mels: 64 |
|
|
|
optimizer_conf: |
|
|
|
optimizer: 'Adam' |
|
|
|
learning_rate: 0.001 |
|
weight_decay: 1e-6 |
|
|
|
model_conf: |
|
|
|
pooling_type: 'ASP' |
|
|
|
train_conf: |
|
|
|
max_epoch: 30 |
|
log_interval: 100 |
|
|
|
|
|
use_model: 'EcapaTdnn' |
|
|
|
use_loss: 'AAMLoss' |