当前位置: 首页 > 工具软件 > ASRT > 使用案例 >

DeepSpeech(tensorflow)和ASRT_SpeechRecognition识别效果对比

皇甫卓君
2023-12-01

######DeepSpeech(tensorflow)######
pip3 install deepspeech
wget https://github.com/mozilla/DeepSpeech/releases/download/v0.9.3/deepspeech-0.9.3-models-zh-CN.pbmm
wget https://github.com/mozilla/DeepSpeech/releases/download/v0.9.3/deepspeech-0.9.3-models-zh-CN.scorer
deepspeech --model deepspeech-0.9.3-models-zh-CN.pbmm --scorer deepspeech-0.9.3-models-zh-CN.scorer --audio t.wav
参考:
https://github.com/mozilla/DeepSpeech
https://deepspeech.readthedocs.io/en/latest/?badge=latest


######ASRT_SpeechRecognition######
https://github.com/nl8590687/ASRT_SpeechRecognition/releases/download/v0.6.1/ASRT_v0.6.1.zip
unzip ASRT_v0.6.1.zip
cd ASRT_v0.6.1
#修改wav语音文件的路径,wav文件要求sampling rate为16000,可用命令ffmpeg -i t.mp3 -ar 16000 -y t.wav修改
vim test.py
python test.py
参考:https://github.com/nl8590687/ASRT_SpeechRecognition

######识别结果######
wav文件中我的语音:“我说一句话,你给我识别一下看看,看效果怎么样”
ASRT_SpeechRecognition识别结果:“我说西化你部西干尔看长他是我怎没”
deepspeech识别结果:“在线”

 类似资料: