MNIST-GAN
Repo consists of a jupyter notebook explaning a simple implementation of a DCGAN model to generate MNIST digits dataset. The GAN model is trained for 100 epochs and has generated images of new images of digits. Keras with TensorFlow as backend has been used for implementation. Snapshot of Digits generated during 80th epoch and 100th epoch are included for reference in this repo.
Note: Using a GPU shall reduce the ETA for completion of GAN model training.