HDCGAN Training 1024x1024 233 epochs - Pittsburgh, Orlando and Manhattan

Описание: J. D. Curtó ([email protected]) ^1,2,3
I. C. Zarza ([email protected]) ^1,2,3

1. Department of Computer Science and Engineering. The Chinese University of Hong Kong.
2. Robotics. Carnegie Mellon.
3. Department of Electrical Engineering. City University of Hong Kong.

HDCGAN (High-resolution Deep Convolutional Generative Adversarial Networks) image training using 2 x NVIDIA Titan X. 233 epochs of the training procedure, using 1024x1024 images from Street View and a batch size of 32, are depicted.