DCGAN implemetation for custom dataset. Inspired from Udacity deep learning nano degree programme.
- Keep you images in "data" folder
- Run train.py script to start training the dcgan
python train.py
Model will resize all the images to 32*32. Checkpoints will be stored in "checkpoints" folder.
- Run test.py
python test.py
This will take load the checkpoint from checkpoints folder and generate random images.
References: Full paper here: http://arxiv.org/abs/1511.06434