Why CNNs detect spatial patterns in PyTorch - Test Your Understanding

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a 2D convolutional layer with 3 input channels and 16 output channels.

PyTorch

conv_layer = torch.nn.Conv2d([1], 16, kernel_size=3, stride=1, padding=1)

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

A1

B16

C32

D3

Attempts:

3 left

💡 Hint

Common Mistakes

Using output channels instead of input channels

Confusing kernel size with channels

The first argument to Conv2d is the number of input channels. For example, an RGB image has 3 channels.

2fill in blank

medium

Complete the code to apply the convolutional layer to the input tensor.

PyTorch

output = conv_layer([1])

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

Aconv_layer

Binput_tensor

Coutput

Dtorch.nn

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the layer itself instead of the input

Passing the output variable before assignment

The convolutional layer is applied to the input tensor to produce the output feature map.

3fill in blank

hard

Fix the error in the code to correctly initialize the convolutional layer with a 5x5 kernel.

PyTorch

conv_layer = torch.nn.Conv2d(3, 16, kernel_size=[1], stride=1, padding=2)

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

A5

B7

C1

D3

Attempts:

3 left

💡 Hint

Common Mistakes

Using kernel size 3 instead of 5

Incorrect padding for kernel size

The kernel size should be 5 to create a 5x5 filter. Padding of 2 keeps the output size same as input.

4fill in blank

hard

Fill both blanks to create a feature map by applying a convolution and then a ReLU activation.

PyTorch

conv_layer = torch.nn.Conv2d(3, 8, kernel_size=3, padding=1)
output = conv_layer([1])
activated_output = torch.nn.functional.[2](output)

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

Ainput_tensor

Bsigmoid

Crelu

Doutput_tensor

Attempts:

3 left

💡 Hint

Common Mistakes

Using output_tensor instead of input_tensor

Using sigmoid instead of relu

The input tensor is passed to the convolutional layer, then ReLU activation is applied to introduce non-linearity.

5fill in blank

hard

Fill all three blanks to create a simple CNN block: convolution, activation, and max pooling.

PyTorch

conv = torch.nn.Conv2d(1, [1], kernel_size=3, padding=1)
x = conv([2])
x = torch.nn.functional.[3](x)
x = torch.nn.functional.max_pool2d(x, 2)

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

A10

Binput_image

Crelu

D5

Attempts:

3 left

💡 Hint

Common Mistakes

Using too many output channels

Passing wrong variable as input

Using wrong activation function

The convolution outputs 5 channels, input_image is the input tensor, and ReLU activation is applied before max pooling.