Loss functions (MSELoss, CrossEntropyLoss) in PyTorch - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a Mean Squared Error loss function in PyTorch.

PyTorch

import torch.nn as nn
loss_fn = nn.[1]()

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

ABCELoss

BMSELoss

CL1Loss

DCrossEntropyLoss

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing CrossEntropyLoss which is for classification tasks.

Using BCELoss which is for binary classification.

The Mean Squared Error loss function in PyTorch is nn.MSELoss(). It measures the average squared difference between predicted and true values.

2fill in blank

medium

Complete the code to create a Cross Entropy loss function in PyTorch.

PyTorch

import torch.nn as nn
loss_fn = nn.[1]()

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

AL1Loss

BMSELoss

CCrossEntropyLoss

DNLLLoss

Attempts:

3 left

💡 Hint

Common Mistakes

Using MSELoss for classification tasks.

Using NLLLoss directly without applying LogSoftmax.

The Cross Entropy loss function in PyTorch is nn.CrossEntropyLoss(). It is used for multi-class classification problems.

3fill in blank

hard

Fix the error in the code to compute the loss between predictions and targets using MSELoss.

PyTorch

import torch
import torch.nn as nn

predictions = torch.tensor([2.5, 0.0, 2.1])
targets = torch.tensor([3.0, -0.5, 2.0])

loss_fn = nn.MSELoss()
loss = loss_fn([1], targets)
print(loss.item())

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

Atargets.detach()

Btargets

Cpredictions.detach()

Dpredictions

Attempts:

3 left

💡 Hint

Common Mistakes

Swapping predictions and targets in the loss function.

Using detached tensors which breaks gradient flow.

The first argument to the loss function should be the model's predictions. Here, predictions is correct.

4fill in blank

hard

Fill both blanks to compute Cross Entropy loss for classification outputs and labels.

PyTorch

import torch
import torch.nn as nn

outputs = torch.tensor([[1.0, 2.0, 0.5], [0.5, 1.5, 1.0]])
labels = torch.tensor([1, [1]])

loss_fn = nn.CrossEntropyLoss()
loss = loss_fn(outputs, [2])
print(loss.item())

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

A0

Blabels

C2

D3

Attempts:

3 left

💡 Hint

Common Mistakes

Using label values outside the range of classes.

Passing outputs as labels or vice versa.

The labels tensor contains class indices. The second label is class 2. The loss function takes outputs and labels as arguments.

5fill in blank

hard

Fill all three blanks to create MSE loss and compute it between predictions and targets.

PyTorch

import torch
import torch.nn as nn

predictions = torch.tensor([[1], [2], 2.0])
targets = torch.tensor([3.0, 0.0, [3]])

loss_fn = nn.MSELoss()
loss = loss_fn(predictions, targets)
print(loss.item())

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

A2.5

B0.0

C2.1

D3.0

Attempts:

3 left

💡 Hint

Common Mistakes

Mixing up prediction and target values.

Using values not consistent with the example.

The predictions tensor has values 2.5, 0.0, and 2.0. The targets tensor has 3.0, 0.0, and 2.1. These values match the example for computing MSE loss.