/1D-GAN

Implementation of a one-dimensional Generative Adversarial Network that generates a cubic function

Primary LanguageJupyter NotebookMIT LicenseMIT

1D-GAN

Implementation of a one-dimensional Generative Adversarial Network that generates a cubic function using Keras.

Results

After 2000 epochs

After 4000 epochs

After 6000 epochs

After 8000 epochs

After 10000 epochs

Improvements:

  1. More number of hidden layers
  2. Different activation functions