February 2018
Intermediate to advanced
450 pages
11h 27m
English
We can fine-tune the GANs by changing the following set of hyperparameters:
# size of discriminator input image#28 by 28 will flattened to be 784input_img_size = 784 # size of the generator latent vectorgen_z_size = 100# number of hidden units for the generator and discriminator hidden layersgen_hidden_size = 128disc_hidden_size = 128#leaky ReLU alpha parameter which controls the leak of the functionleaky_relu_alpha = 0.01# smoothness of the label label_smooth = 0.1
Read now
Unlock full access