/ASR_dosmono

Automatic Speech Recognition with TensorFlow(CNN+BLSTM+CTC)

Primary LanguagePython

中文语音识别

1. 环境
  • Python：3.6.5
  • Tensorflow : 1.8.0
2. 训练数据下载
3. 训练

  • 配置conf.ini [FILE_DATA] 2 wav_path=/home/speech.AI/github/speech_recognition/opt/wav/test/ 3 label_file=/home/speech.AI/github/speech_recognition/opt/wav/test.word.txt 4 savedir=/home/speech.AI/github/speech_recognition/opt/tf/speechtest/ 5 savefile=speech.cpkt 6 tensorboardfile=/home/speech.AI/github/speech_recognition/opt/tf/logs

  • 在终端运行 python train.py 开始训练

  • 在终端运行 python test.py 测试

  • 也可以使用PyCharm打开

4. 测试效果

test-image