Variational Autoencoder

Reconstruction:

Hyperparameters:

• learning_rate=0.005
• training_epochs=15
• Batch Size =100

Network Structure:

Number of Neurons:

• Encoder:
  • Hidden layer 1: 600
  • Hidden layer 2: 500
• Decoder:
  • Hidden layer 1: 600
  • Hidden layer 2: 500

Latent Space: 15d

Execution:

Output:

2d latent space:

Hyperparameters:

• learning_rate=0.005
• training_epochs=15
• Batch Size =5000

Network Structure:

Number of Neurons:

• Encoder:
  • Hidden layer 1: 600
  • Hidden layer 2: 500
• Decoder:
  • Hidden layer 1: 600
  • Hidden layer 2: 500

Latent Space: 2d

Execution:

Output:

Plot: