深入浅出语音识别-yesno例子剖析

源文件如下：

#！/bin/bash

train_cmd="utils/run.pl"

decode_cmd="utils/run.pl"

#========== 下载原始数据文件并解压到waves_yesno目录========

if [! -d waves_yesno ]; then
   wget http://www.openslr.org/resources/1/waves_yesno.tar.gz || exit 1;
   tar -xvzf waves_yesno.tar.gz || exit 1;
fi

train_yesno=train_yesno
test_base_name=test_yesno

#========== 删除上一次运行生成的文件==================
rm -rf data exp mfcc

#======== 第一步：数据准备 =========================
local/prepare_data.sh waves_yesno
local/prepare_dict.sh
utils/prepare_lang.sh --position-dependent-phones false data/local/dict "<SIL>" data/local/lang data/lang
local/prepare_lm.sh

#========第二步：语音信号特征抽取 ====================
for x in train_yesno test_yesno; do
  steps/make_mfcc.sh --nj 1 data/$x exp/make_mfcc/$x mfcc
  steps/compute_cmvn_stats.sh data/$x exp/make_mfcc/$x mfcc
  utils/fix_data_dir.sh data/$x
done

#========第三步：单音模型训练 =======================
steps/train_mono.sh --nj 1 --cmd "$train_cmd" --totgauss 400 data/train_yesno data/lang exp/mono0a

#========第四步：生成解码矩阵 =======================
utils/mkgraph.sh data/lang_test_tg exp/mono0a exp/mono0a/graph_tgpr

#========第五步：解码 =============================
steps/decode.sh --nj 1 --cmd "$decode_cmd" exp/mono0a/graph_tgpr data/test_yesno exp/mono0a/decode_test_yesno
for x in exp/*/decode*; do
  [ -d $x ] && grep WER $x/wer_* | utils/best_wer.sh;
done

########################################################################

下面再对每一步进行详细