Deep Learning with PyTorch

Creating a fully connected model and train

We will use a simple linear model, similar to what we used in ResNet and Inception. The following code shows the network architecture which we will be using to train the model:

class FullyConnectedModel(nn.Module):        def __init__(self,in_size,out_size):        super().__init__()        self.fc = nn.Linear(in_size,out_size)    def forward(self,inp):        out = self.fc(inp)        return outfc = FullyConnectedModel(fc_in_size,classes)if is_cuda:    fc = fc.cuda()

We will use the same fit method to train the preceding model. The following code snippet shows the training code, along with the results:

train_losses , train_accuracy = [],[]val_losses , val_accuracy = [],[]for epoch in range(1,10): epoch_loss, epoch_accuracy = fit(epoch,fc,trn_feat_loader,phase='training') ...