PyTorch-Models

• Implementing PyTorch models as seen in the below figures

Sequential 1: Model with no hidden layer

• Train for 10 epochs (output is softmax)

  

Sequential 2: Model with single hidden layer

• Use sigmoid in hidden layer, train for 20 epochs

  

  Sequential 3: Deep model

• Use ReLU in hidden layer, train for 20 epochs

  

  Functional: Complex deep model

• Use sigmoid for layers within the blue box and use ReLU in the remaining hidden layers, train for 25 epochs