search

osh / KerasGAN

A couple of simple GANs in Keras
501 stars 177 forks source link

Generative network output shape #11

Open fasilpv opened 7 years ago

fasilpv commented 7 years ago

Why is the generative network output shape is (None, 400, 28, 1). Should not this be (None, 28, 28, 1)? I am referring mnist_gan.py

rafael-glima commented 7 years ago

Try using data_format="channels_first" in the Convolution2D and Upsampling2D layers