What should I do if I want to use this to train my own database about generative-adversarial-networks HOT 2 CLOSED

You'll need to replace the lines where we call the function mnist.train.next_batch. For example, in the line

real_image_batch = mnist.train.next_batch(batch_size)[0].reshape([batch_size, 28, 28, 1])

you'll need to create your own python function that basically returns a random subset of your dataset. For example, if you have 25,000 images that are each 28x28, you should have a 25000 x 28 x 28 matrix that represents those examples. You'll have to create a function that selects a random subset of those. The shape of the matrix you return will be batch_size x 28 x 28.

from generative-adversarial-networks.

Thanks Adit!

from generative-adversarial-networks.