def get_output_for(self, input, **kwargs):
if input.ndim > 2:
# if the input has more than two dimensions, flatten it into a
# batch of feature vectors.
input = T.flatten(input, 2)
activation = T.dot(input, self.W)
if self.b is not None:
activation = T.broadcast('+', activation , T.dimshuffle(self.b, 'x', 0), 'xx,1x')
return self.nonlinearity(activation)
def get_output_for(self, input, **kwargs):
conved = self.convolve(input, **kwargs)
if self.b is None:
activation = conved
elif self.untie_biases:
activation = conved + T.shape_padleft(self.b, 1)
else:
activation = conved + T.dimshuffle(self.b, ('x', 0) + ('x',) * self.n)
return self.nonlinearity(activation)
