Skip to content

Activation function for discriminator of WGAN-GP #16

@yutaochen1998

Description

@yutaochen1998

In the network architecture of discriminator:
tf.keras.layers.Dense(units=1, activation="sigmoid"),
You don't need any activation function for discriminator, since you are using Wasserstein Loss. Using sigmoid here would greatly limit your convergeance speed.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions