Dataset class (custom datasets) in PyTorch - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of a simple custom Dataset __getitem__

What is the output of this code snippet when accessing dataset[1]?

PyTorch

import torch
from torch.utils.data import Dataset

class MyDataset(Dataset):
    def __init__(self):
        self.data = [10, 20, 30]
    def __len__(self):
        return len(self.data)
    def __getitem__(self, idx):
        return self.data[idx] * 2

dataset = MyDataset()
print(dataset[1])

A 40

B IndexError

C 30

D 20

💡 Hint

Remember __getitem__ returns data[idx] multiplied by 2.

The __getitem__ method returns the element at index 1 (which is 20) multiplied by 2, so 40.

❓ data_output

intermediate

Length of a custom Dataset

What is the output of len(dataset) for this custom Dataset?

PyTorch

from torch.utils.data import Dataset

class SampleDataset(Dataset):
    def __init__(self):
        self.samples = list(range(7))
    def __len__(self):
        return len(self.samples)
    def __getitem__(self, idx):
        return self.samples[idx]

dataset = SampleDataset()
print(len(dataset))

A 6

B 7

C TypeError

D 8

💡 Hint

Check how many elements are in samples list.

The samples list has 7 elements (0 to 6), so len(dataset) returns 7.

🔧 Debug

advanced

Identify the error in this custom Dataset code

What error will this code raise when trying to create the dataset instance?

PyTorch

from torch.utils.data import Dataset

class BrokenDataset(Dataset):
    def __init__(self):
        self.data = [1, 2, 3]
    def __getitem__(self, idx):
        return self.data[idx]

# Missing __len__ method

dataset = BrokenDataset()
print(len(dataset))

A TypeError: object of type 'BrokenDataset' has no len()

B AttributeError: 'BrokenDataset' object has no attribute '__len__'

C No error, dataset created successfully

D IndexError when accessing dataset[0]

💡 Hint

Dataset requires __len__ method to work properly.

PyTorch Dataset requires __len__ method. Without it, calling len(dataset) raises TypeError.

🚀 Application

advanced

Custom Dataset with transform application

Given this custom Dataset, what will be the output of dataset[0]?

PyTorch

from torch.utils.data import Dataset

def add_five(x):
    return x + 5

class TransformDataset(Dataset):
    def __init__(self, data, transform=None):
        self.data = data
        self.transform = transform
    def __len__(self):
        return len(self.data)
    def __getitem__(self, idx):
        sample = self.data[idx]
        if self.transform:
            sample = self.transform(sample)
        return sample

data = [1, 2, 3]
dataset = TransformDataset(data, transform=add_five)
print(dataset[0])

A 1

B 5

C TypeError

D 6

💡 Hint

Transform adds 5 to the original data element.

dataset[0] returns data[0] + 5 = 1 + 5 = 6 because transform is applied.

🧠 Conceptual

expert

Why implement a custom Dataset class in PyTorch?

Which of the following is the best reason to create a custom Dataset class in PyTorch?

A To replace the need for DataLoader in batching data

B To automatically parallelize model training across GPUs

C To define how to load and access data samples and their labels efficiently

D To speed up the training by compiling the model

💡 Hint

Think about what Dataset controls in PyTorch data handling.

Custom Dataset defines how data is loaded and accessed, enabling flexible data handling.