Why nn.Module organizes model code in PyTorch - Test Your Understanding

Practice - 5 Tasks

1fill in blank

easy

Complete the code to define a PyTorch model class that inherits from nn.Module.

PyTorch

import torch.nn as nn

class SimpleModel(nn.Module):
    def __init__(self):
        super(SimpleModel, self).__init__()
        self.linear = nn.Linear(10, 1)

    def [1](self, x):
        return self.linear(x)

Drag options to blanks, or click blank then click option'

Aforward

Bapply_activation

Cbackward

Dtrain

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'backward' or 'train' instead of 'forward' for the method name.

The forward method defines how the input data passes through the model layers.

2fill in blank

medium

Complete the code to initialize the parent class in the model constructor.

PyTorch

class MyModel(nn.Module):
    def __init__(self):
        [1]
        self.layer = nn.Linear(5, 2)

Drag options to blanks, or click blank then click option'

Asuper().__init__()

Bself.__init__()

Cnn.Module.__init__(self)

Dsuper(MyModel, self).__init__()

Attempts:

3 left

💡 Hint

Common Mistakes

Calling self.__init__() causes infinite recursion.

Using super().__init__() without arguments is valid in Python 3 but here we want the explicit form.

Calling super(MyModel, self).__init__() properly initializes the parent nn.Module class.

3fill in blank

hard

Fix the error in the forward method to correctly use the linear layer.

PyTorch

class FixModel(nn.Module):
    def __init__(self):
        super(FixModel, self).__init__()
        self.linear = nn.Linear(3, 1)

    def forward(self, x):
        return self.linear[1](x)

Drag options to blanks, or click blank then click option'

A.call

B.forward

C()

D.apply

Attempts:

3 left

💡 Hint

Common Mistakes

Trying to call '.forward(x)' directly instead of using parentheses.

In PyTorch, layers are called like functions with parentheses, e.g., self.linear(x).

4fill in blank

hard

Fill both blanks to create a model with two layers and use ReLU activation between them.

PyTorch

class TwoLayerModel(nn.Module):
    def __init__(self):
        super(TwoLayerModel, self).__init__()
        self.fc1 = nn.Linear(4, 3)
        self.fc2 = nn.Linear(3, 1)
        self.relu = nn.ReLU()

    def forward(self, x):
        x = self.fc1[1](x)
        x = self.relu[2](x)
        return self.fc2(x)

Drag options to blanks, or click blank then click option'

A()

B.forward

C.call

D.apply

Attempts:

3 left

💡 Hint

Common Mistakes

Using '.forward' or '.call' instead of parentheses to call layers.

Layers and activations are called with parentheses like functions, e.g., self.fc1(x) and self.relu(x).

5fill in blank

hard

Fill all three blanks to define a model that stores layers as attributes, calls them in forward, and returns the output.

PyTorch

class CustomModel(nn.Module):
    def __init__(self):
        super(CustomModel, self).__init__()
        self.layer1 = nn.Linear(6, 4)
        self.activation = nn.Sigmoid()
        self.layer2 = nn.Linear(4, 2)

    def forward(self, input):
        out = self.layer1[1](input)
        out = self.activation[2](out)
        return self.layer2[3](out)

Drag options to blanks, or click blank then click option'

A()

B.forward

C.call

D.apply

Attempts:

3 left

💡 Hint

Common Mistakes

Trying to call layers with '.forward' or '.call' instead of parentheses.

Each layer and activation is called like a function using parentheses, e.g., self.layer1(input).