Open wwfcnu opened 11 months ago
想问一下,我看model目录下面一些demo用到的接口是短语音识别(1分钟以内),如果我的测试语音超过了1分钟,是不是就要用录音文件转写的接口了
是的,我们的测试集都是切好的短句。测试长音频的话,现在model/下的模型不保证都支持。
想问一下,我看model目录下面一些demo用到的接口是短语音识别(1分钟以内),如果我的测试语音超过了1分钟,是不是就要用录音文件转写的接口了