|
|
|
|
# https://yaml.org/type/float.html
|
|
|
|
|
###########################################
|
|
|
|
|
# Data #
|
|
|
|
|
###########################################
|
|
|
|
|
train_manifest: data/manifest.train
|
|
|
|
|
dev_manifest: data/manifest.dev-clean
|
|
|
|
|
test_manifest: data/manifest.test-clean
|
|
|
|
|
min_input_len: 0.0
|
|
|
|
|
max_input_len: 30.0 # second
|
|
|
|
|
min_output_len: 0.0
|
|
|
|
|
max_output_len: .inf
|
|
|
|
|
min_output_input_ratio: 0.00
|
|
|
|
|
max_output_input_ratio: .inf
|
|
|
|
|
|
|
|
|
|
###########################################
|
|
|
|
|
# Dataloader #
|
|
|
|
|
###########################################
|
|
|
|
|
batch_size: 15
|
|
|
|
|
mean_std_filepath: data/mean_std.json
|
|
|
|
|
unit_type: char
|
|
|
|
|
vocab_filepath: data/lang_char/vocab.txt
|
|
|
|
|
augmentation_config: conf/augmentation.json
|
|
|
|
|
random_seed: 0
|
|
|
|
|
spm_model_prefix:
|
|
|
|
|
spectrum_type: linear
|
|
|
|
|
feat_dim:
|
|
|
|
|
target_sample_rate: 16000
|
|
|
|
|
max_freq: None
|
|
|
|
|
n_fft: None
|
|
|
|
|
stride_ms: 10.0
|
|
|
|
|
window_ms: 20.0
|
|
|
|
|
delta_delta: False
|
|
|
|
|
dither: 1.0
|
|
|
|
|
use_dB_normalization: True
|
|
|
|
|
target_dB: -20
|
|
|
|
|
random_seed: 0
|
|
|
|
|
keep_transcription_text: False
|
|
|
|
|
sortagrad: True
|
|
|
|
|
shuffle_method: batch_shuffle
|
|
|
|
|
num_workers: 0
|
|
|
|
|
|
|
|
|
|
############################################
|
|
|
|
|
# Network Architecture #
|
|
|
|
|
############################################
|
|
|
|
|
num_conv_layers: 2
|
|
|
|
|
num_rnn_layers: 3
|
|
|
|
|
rnn_layer_size: 2048
|
|
|
|
|
rnn_direction: forward
|
|
|
|
|
num_fc_layers: 2
|
|
|
|
|
fc_layers_size_list: 512, 256
|
|
|
|
|
use_gru: False
|
|
|
|
|
blank_id: 0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
###########################################
|
|
|
|
|
# Training #
|
|
|
|
|
###########################################
|
|
|
|
|
n_epoch: 50
|
|
|
|
|
accum_grad: 4
|
|
|
|
|
lr: 1.0e-3
|
|
|
|
|
lr_decay: 0.83
|
|
|
|
|
weight_decay: 1.0e-6
|
|
|
|
|
global_grad_clip: 5.0
|
|
|
|
|
log_interval: 100
|
|
|
|
|
checkpoint:
|
|
|
|
|
kbest_n: 50
|
|
|
|
|
latest_n: 5
|
