aserdega / convlstmgru Goto Github PK

Thank you for offering great code:)

I have 2 questions regarding batchnormalization in your code.
As shown in convlstm.py's batchnorm part like below

def reset_parameters(self):
#self.conv.reset_parameters()
nn.init.xavier_uniform_(self.conv.weight, gain=nn.init.calculate_gain('tanh'))
self.conv.bias.data.zero_()

  if self.batchnorm:
      self.bn1.reset_parameters()
      self.bn2.reset_parameters()

Q1) Does it mean that batchnorm is operating between (vertically) convlstm layer?
or between (horizontally) convlstm cell?

Q2) Is it really work as batchnormalization? As I know of, xavier uniform is not exactly same as batchnormalization.

My question might be wrong, thus waiting for your merciful answer:)