An implementation of Infogan [1] for pytorch. For an explanation of the project see [2].
You can train a new model using python trainer.py.
Some pretrained models are included with the code for inference. You can pass them to the network using the --modelpath argument in test.py.
You can test a newly trained model or one of the pretrained moedls with python test.py --expname EXPNAME --modelpath "PATH_TO_PRETRAINED_MODELS" --modelname "NAME_OF_GENERATOR".
[1] - X. Chen, Y. Duan, R. Houthooft, J. Schulman, I. Sutskever, P. Abbeel - InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets
[2] - Deepankar C. - Generation of Images via Attribute Manipulation using Disentangled Representation Learning