Conditional Wasserstein GAN is a project assignment for COMP9501, 2022 in The University of Hong Kong. The application is for the renewable energy scenario generation.
To launch training process, run
pyhon train.py
The trained generator and discriminator will be saved as loggings/{pv, wind}_C{True, False}_{G, D}.pth
.
In this name,
{pv, wind}
represents the type of renewable energy power: solar or wind;{True, False}
represents whether have conditions;{G, D}
represents generator/discriminator.
To change the hyperparameters, please check train.py
for training parameters and gan.py
for model structures.
To have the visual results, please check test.ipynb
Chenxi Wang and Dalin Qin