nn.Conv1d当kernel-size=1时的应用

Conv1d和Conv2d的区别

Conv1d的官方文档

重点就是理解Conv1d纵列进行滑动

In [12]: a = torch.randn(20,16,50)

In [13]: a.shape
Out[13]: torch.Size([20, 16, 50])

In [14]: m = torch.nn.Conv1d(16, 33, 1)

In [15]: m(a).shape
Out[15]: torch.Size([20, 33, 50])