Conditional version of WGAN-GP is the combination of cgan and wgan-gp. It can be used to generate samples of a particular class.
Mnist dataset is used in the experiment.
python3 cwgan_gp.py
Use the code wgan.generate_images(class_names).
Need to change img_rows, img_cols, channels, nclasses, load your own dataset and change the network structure to corresponding size.