Deep-Learning-for-GANs

1. Upsampling with Convolutional Neural Networks

• Using Upsampling Layer
• Using the Transpose Convolutional Layer

2. Implementing GAN Hacks to Train Stable Models

• Deep Convolutional GANs (DCGANs)
  • Downsample Using Strided Convolutions
  • Upsample Using Strided Convolutions
  • Using Leaky ReLU
  • Using Batch Normalization
  • Using Gaussian Weight Initialization
  • Using Adam Stochastic Gradient Descent
  • Scaling images to the Range[-1,1]
• Soumith Chintala's GAN Hacks
  • Using a Gaussian Latent Space
  • Using Label Smoothing
  • Using Noisy Labels

3. Developing a 1D GAN from Scratch

• Selecting a One-Dimensional Function
• Defining a Discriminator Model
• Defining a Generator Model
• Training the Generator Model
• Evaluating the Performance of the GAN
• Complete Example of Training the GAN