kaldi-gstreamer-server docker环境搭建
https://github.com/jcsilva/docker-kaldi-gstreamer-server
Pull Docker Hub
1
| docker pull jcsilva/docker-kaldi-gstreamer-server
|
准备模型以及yaml配置文件,此处demo搭建使用的是0-10的英文数字识别的模型
1
2
| [root@centos_asr kaldi_models]$ ls /root/xin/kaldi_models/digits
final.mdl HCLG.fst words.txt
|
1
2
3
4
5
6
7
8
9
10
11
| [root@centos_asr kaldi_models]$ head digits_tri1.yaml
timeout-decoder : 10
decoder:
model: xin/kaldi_models/digits/final.mdl
word-syms: xin/kaldi_models/digits/words.txt
fst: xin/kaldi_models/digits/HCLG.fst
silence-phones: "1:2:3:4:5"
out-dir: tmp
use-vad: False
silence-timeout: 60
|
运行docker环境，并把本地模型数据暴露给docker
1
| docker run -it -p 8080:80 -v /root/xin/kaldi_models:/opt/xin/kaldi_models jcsilva/docker-kaldi-gstreamer-server:latest /bin/bash
|
在container中启动服务
1
| /opt/start.sh -y /opt/xin/kaldi_models/digits_tri1.yaml
|
服务启动成功之后即可用于识别准备好的音频：
1
2
3
| root@cd4dcab22d74:/opt# python kaldi-gstreamer-server/kaldigstserver/client.py -u ws://1.14.104.94:8080/client/ws/speech -r 32000 xin/kaldi_models/1-789.wav
seven eight nine.
|