Generative Adversarial Networks Projects

Tuning the hyperparameters

The hyperparameters are the properties of a model, which are fixed during the training of the model. Different parameters can have different accuracy. Let's take a look at some of the common hyperparameters used:

• Learning rate
• Batch size
• Number of epochs
• Generator optimizer
• Discriminator optimizer
• Number of layers
• Number of units in a dense layer
• Activation function
• Loss function

In the Implementing a DCGAN using Keras section, the learning rates were fixed: 0.0005 for the generator model and 0.0005 for the discriminator model. The batch size was 128. Tweaking these values might lead us to create a better model. If your model is not generating plausible images, try changing these values and run your model again.  ...