array dimension not consistent error

[qigangliu]

this project is great, i just search for a CycleGAN implemented in keras, when i run the demo, i encounter a error "expected input_3 to have shape (None, 3, 128, 128) but got array with shape (1, 128, 128, 3)" accured in "fake_A_pool.extend(self.BtoA.predict(real_B))" , it seems there are some problems with the array dimensions conversion, can you confirm it, thank you very much.

[Shaofanl]

Indeed. There is a difference between TF and Theano in dimension. I think I used "image_data_format": "channel_last" in $HOME/.keras/keras.json when I wrote this repository. Since they stop to update Theano anymore, I recommend you to change everything to TF-like yourself.