Tensorflow implementation of (https://arxiv.org/pdf/1703.10593.pdf).
Cycle Consistent GANs are an adaptation of Generative Adversarial Networks, in which the resulting model has the capability of performing domain adaptation between two datasets of varying domains. Again- unpaired! Images between datasets don't need to be directly matched, as the additional cycle consistency term added within the CycleGAN model allows for additional stability within training - of which pressures output domains to be consistent.
Sample mappings shown above.
Dataset available at https://github.com/junyanz/CycleGAN (horse2zebra preferred, else change image dimensions). Alter path names in main for local directory for proper usage.
Packages Required in Environment:
- Tensorflow
- CV2
- Numpy
- Matplotlib
GPU training is preferred.
To execute the program, use the following command whilst in terminal:
python main.py