Python notebook containing TensorFlow DCGAN implementation. It was trained on a Celebrities dataset.
Check out corresponding Medium article: Face Generator - Generating Artificial Faces with Machine Learning ๐ง.
Check out corresponding Kaggle kernel: Face Generator.
Network architecture by Radford et al., 2015.
Visualization of training with the following hyperparameteres.
DATASET_SIZE = 100000
IMAGE_SIZE = 128
NOISE_SIZE = 100
LR_D = 0.00004
LR_G = 0.0002
BATCH_SIZE = 64
EPOCHS = 20
BETA1 = 0.5
WEIGHT_INIT_STDDEV = 0.02
EPSILON = 0.00005
Here considering the computational power, we will be running only 20 epochs ratehr than 60 epochs. Instead you can view the result after running 60 epochs which has been ran on a different machine with higher computational capability.
Generated samples after 60 epochs of training.