Daniel Voigt Godoy - Deep Learning with PyTorch Step-by-Step A Beginner’s Guide-leanpub

Model Configuration
We can easily translate the model depicted above to code:
Model Configuration
1 # Sets learning rate - this is "eta" ~ the "n"-like Greek letter
2 lr = 0.1
3
4 torch.manual_seed(17)
5 # Now we can create a model
6 model_nn = nn.Sequential()
7 model_nn.add_module('flatten', nn.Flatten())
8 model_nn.add_module('hidden0', nn.Linear(25, 5, bias=False))
9 model_nn.add_module('hidden1', nn.Linear(5, 3, bias=False))
10 model_nn.add_module('output', nn.Linear(3, 1, bias=False))
11 model_nn.add_module('sigmoid', nn.Sigmoid())
12
13 # Defines an SGD optimizer to update the parameters
14 optimizer_nn = optim.SGD(model_nn.parameters(), lr=lr)
15
16 # Defines a binary cross-entropy loss function
17 binary_loss_fn = nn.BCELoss()
I’ve kept the names of the modules consistent with the captions in the figure so it is
easier to follow. The rest of the code should already be familiar to you.
Model Training
Let’s train our new deep-ish model for 100 epochs using the StepByStep class and
visualize the losses:
Model Training
1 n_epochs = 100
2
3 sbs_nn = StepByStep(model_nn, binary_loss_fn, optimizer_nn)
4 sbs_nn.set_loaders(train_loader, val_loader)
5 sbs_nn.train(n_epochs)
304 | Chapter 4: Classifying Images