README.md

DCGAN

dcgan_c.py - production (composable)

Paper

Macro-Architecture

Generator

Discriminator

Composable

Example: Compose and Train for CIFAR-30

gan = DCGAN(input_shape=(32, 32, 3))
gan.model.summary()

from tensorflow.keras.datasets import cifar10
import numpy as np
(x_train, _), (_, _) = cifar10.load_data()
x_train = x_train / 127.5 - 1.
gan.train(x_train, latent=100, epochs=6000)