Shiki

This caused by the different version of Keras. You could do the following changes. ``` def discriminator_model(): model = Sequential() model.add(Convolution2D( 64, 5, 5, border_mode='same', input_shape=(28, 28, 1))) #modify this...

The default parameter of 'data_format' for Conv2d layer and pooling layer is 'channel_last', [see \[Keras doc\]](https://keras.io/layers/convolutional/) While the input data format is 'channel_first', it causes the conflict. You can either...