实验1：基于DNN-HMM结构的语音识别系统

最新推荐文章于 2023-01-10 23:49:01 发布

猪猪小宝y

最新推荐文章于 2023-01-10 23:49:01 发布

阅读量827

点赞数 1

CC 4.0 BY-SA版权

分类专栏：语音识别基本法文章标签： nlp

本文链接：https://blog.youkuaiyun.com/qq_37443534/article/details/119042314

基于Kaldi的DNN-HMM语音识别系统，run.sh文件包含从前期数据准备到最后解码的整个过程，该脚本是语音识别各个步骤的封装。

#!/bin/dash
#bash or dash
# Copyright 2016  Tsinghua University (Author: Dong Wang, Xuewei Zhang)
#           2018  Tsinghua University (Author: Zhiyuan Tang)
# Apache 2.0.

. ./cmd.sh ## You'll want to change cmd.sh to something that will work on your system.
           ## This relates to the queue.
           ## 根据系统将cmd.sh最后三行改成queue.pl or run.pl
. ./path.sh


n=8 # parallel jobs

set -euo pipefail # 管道符中任意命令出错或者遇到未定义变量或方法时都停止运行


###### Bookmark: basic preparation ######

# corpus and trans directory
# 指定数据集和训练文件目录
# 数据集用thchs30
thchs=/home/yy/kaldi-trunk/egs/cslt_cases/asr_baseline/data/data_thchs30

# 下载数据集
# you can obtain the database by uncommting the following lines
# [ -d $thchs ] || mkdir -p $thchs 
# echo "downloading THCHS30 at $thchs ..."
# local/download_and_untar.sh $thchs  http://www.openslr.org/resources/18 data_thchs30 
# local/download_and_untar.sh $thchs  http://www.openslr.org/r