The batch number in deformed convolution

[RuihanDou]

I only import deformed_convolution to adopt deformed convolution to my network. The issue is that

self.num_batch = input_shape[0]

Actually the input_shape[0] is None when building the network, the operation liketf.reshape()and tf.tile() can't be done.

There 's error like
TypeError: Failed to convert object of type <class 'list'> to Tensor. Contents: [None, 1, 1, 1]. Consider casting elements to a supported type