|
# network architecture
|
model: Paraformer
|
model_conf:
|
ctc_weight: 0.3
|
lsm_weight: 0.1
|
length_normalized_loss: false
|
predictor_weight: 1.0
|
sampling_ratio: 0.4
|
use_1st_decoder_loss: true
|
|
# encoder
|
encoder: ConformerEncoder
|
encoder_conf:
|
output_size: 256 # dimension of attention
|
attention_heads: 4
|
linear_units: 2048 # the number of units of position-wise feed forward
|
num_blocks: 12 # the number of encoder blocks
|
dropout_rate: 0.1
|
positional_dropout_rate: 0.1
|
attention_dropout_rate: 0.0
|
input_layer: conv2d # encoder architecture type
|
normalize_before: true
|
pos_enc_layer_type: rel_pos
|
selfattention_layer_type: rel_selfattn
|
activation_type: swish
|
macaron_style: true
|
use_cnn_module: true
|
cnn_module_kernel: 15
|
|
# decoder
|
decoder: ParaformerSANDecoder
|
decoder_conf:
|
attention_heads: 4
|
linear_units: 2048
|
num_blocks: 6
|
dropout_rate: 0.1
|
positional_dropout_rate: 0.1
|
self_attention_dropout_rate: 0.0
|
src_attention_dropout_rate: 0.0
|
|
# predictor
|
predictor: CifPredictor
|
predictor_conf:
|
idim: 256
|
threshold: 1.0
|
l_order: 1
|
r_order: 1
|
tail_threshold: 0.45
|
|
# frontend related
|
frontend: WavFrontend
|
frontend_conf:
|
fs: 16000
|
window: hamming
|
n_mels: 80
|
frame_length: 25
|
frame_shift: 10
|
lfr_m: 1
|
lfr_n: 1
|
|
specaug: SpecAug
|
specaug_conf:
|
apply_time_warp: true
|
time_warp_window: 5
|
time_warp_mode: bicubic
|
apply_freq_mask: true
|
freq_mask_width_range:
|
- 0
|
- 30
|
num_freq_mask: 2
|
apply_time_mask: true
|
time_mask_width_range:
|
- 0
|
- 40
|
num_time_mask: 2
|
|
train_conf:
|
accum_grad: 1
|
grad_clip: 5
|
max_epoch: 150
|
keep_nbest_models: 10
|
avg_nbest_model: 10
|
log_interval: 50
|
|
optim: adam
|
optim_conf:
|
lr: 0.0005
|
scheduler: warmuplr
|
scheduler_conf:
|
warmup_steps: 30000
|
|
dataset: AudioDataset
|
dataset_conf:
|
index_ds: IndexDSJsonl
|
batch_sampler: EspnetStyleBatchSampler
|
batch_type: length # example or length
|
batch_size: 25000 # if batch_type is example, batch_size is the numbers of samples; if length, batch_size is source_token_len+target_token_len;
|
max_token_length: 2048 # filter samples if source_token_len+target_token_len > max_token_length,
|
buffer_size: 1024
|
shuffle: True
|
num_workers: 4
|
preprocessor_speech: SpeechPreprocessSpeedPerturb
|
preprocessor_speech_conf:
|
speed_perturb: [0.9, 1.0, 1.1]
|
|
tokenizer: CharTokenizer
|
tokenizer_conf:
|
unk_symbol: <unk>
|
|
ctc_conf:
|
dropout_rate: 0.0
|
ctc_type: builtin
|
reduce: true
|
ignore_nan_grad: true
|
normalize: null
|
