I need to downsample a tensor by a factor 2, it is a 1-D tensor:

I don't know how can I do it

CodePudding user response：

If you want to do it using interpolation, the easiest way to do it would be through average pooling. Learn more here.

CodePudding user response：

What you are you are looking for is probably tf.squeeze(tensor), which removes the dimensions that are 1