Replies: 1 comment 1 reply
-
这个我们应该当时直接忽略了,没有做其他处理,你那边直接忽略就行。 |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
前辈们好,有一条测试数据:TEST_NET_Y0000000004_0ub4ZzdHzBc_S00023 TEST_NET_Y0000000004_0ub4ZzdHzBc 180.65 180.73
这个数据只有80ms,经过(窗口大小25ms,窗移10ms)数据处理提出(1,6,80) 维的特征,经过convsubsampling layer的第一个卷积层(size=3,stride=2)之后变成(1,1,2,19),这个时候再过第二个卷积层(size=3,stride=2),这时特征size(2)就小于kernel的size(3)然后就报错了。我有两个问题,1. 降采样曾是否需要进行padding?2. 小于85ms的音频还送入网络识别吗(85ms是降采样卷积层不padding不报错的最短音频时长)?
Beta Was this translation helpful? Give feedback.
All reactions