@ShoufaChen At the moment you only have one good option: define a `self.shape` 
in your block and add a `set_shape()` function that you call from the training 
top level using data's shape before you pass data into the function. 

[ Full content available at: 
https://github.com/apache/incubator-mxnet/issues/9288 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to