You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
PaddleSpeech/examples/ted_en_zh/st0/conf/transformer.yaml

99 lines
2.7 KiB

# https://yaml.org/type/float.html
###########################################
# Data #
###########################################
train_manifest: data/manifest.train
dev_manifest: data/manifest.dev
test_manifest: data/manifest.test
min_input_len: 0.05 # second
max_input_len: 30.0 # second
min_output_len: 0.0 # tokens
max_output_len: 400.0 # tokens
min_output_input_ratio: 0.01
max_output_input_ratio: 20.0
###########################################
# Dataloader #
###########################################
vocab_filepath: data/lang_char/vocab.txt
unit_type: 'spm'
spm_model_prefix: data/lang_char/bpe_unigram_8000
mean_std_filepath: ""
preprocess_config: conf/preprocess.yaml
batch_size: 16
maxlen_in: 5 # if input length > maxlen-in, batchsize is automatically reduced
maxlen_out: 150 # if output length > maxlen-out, batchsize is automatically reduced
raw_wav: True # use raw_wav or kaldi feature
spectrum_type: fbank #linear, mfcc, fbank
feat_dim: 80
delta_delta: False
dither: 1.0
target_sample_rate: 16000
max_freq: None
n_fft: None
stride_ms: 10.0
window_ms: 25.0
use_dB_normalization: True
target_dB: -20
random_seed: 0
keep_transcription_text: False
sortagrad: True
shuffle_method: batch_shuffle
num_workers: 2
############################################
# Network Architecture #
############################################
cmvn_file: "data/mean_std.json"
cmvn_file_type: "json"
# encoder related
encoder: transformer
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 input type, you can chose conv2d, conv2d6 and conv2d8
normalize_before: true
# decoder related
decoder: transformer
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
# hybrid CTC/attention
model_conf:
asr_weight: 0.0
ctc_weight: 0.0
lsm_weight: 0.1 # label smoothing option
length_normalized_loss: false
###########################################
# Training #
###########################################
n_epoch: 120
accum_grad: 2
global_grad_clip: 5.0
optim: adam
optim_conf:
lr: 2.5
weight_decay: 1.0e-06
scheduler: noam
scheduler_conf:
warmup_steps: 25000
lr_decay: 1.0
log_interval: 50
checkpoint:
kbest_n: 50
latest_n: 5