This is the official code for Text to Image Synthesis using Generative Adversarial Networks. Please be aware that the code is in an experimental stage and it might require some small tweaks.
If you find my research useful, please use the following to cite:
@article{Bodnar2018TextTI,
title={Text to Image Synthesis Using Generative Adversarial Networks},
author={Cristian Bodnar},
journal={CoRR},
year={2018},
volume={abs/1805.00676}
}
As it can be seen, the generated images do not suffer from mode collapse.
Illustration of Conditional Wasserstein Progressive Growing GAN on the flowers dataset:
Samples from the birds dataset
PYTHONPATH
to point to the root directory of the project./data
directory./data/flowers/jpg
. You can alternatively run python preprocess/download_flowers_dataset.py
from the
root directory of the project.python preprocess/preprocess_flowers.py
script from the root directory of the project.