Question bank

Pick an answer to see if you're right — with an explanation.

All topics Python SQL / MySQL Data Science NumPy Pandas Matplotlib Machine Learning Deep Learning Artificial Intelligence Large Language Models Django APIs Git & GitHub Power BI Excel Statistics

Deep LearningEasy

What does calling `loss.backward()` do in PyTorch?

aUpdates the model weights

bComputes gradients of the loss with respect to parameters that require grad

cResets all gradients to zero

dRuns the forward pass again

Deep LearningEasy

Which layer type is best suited to capturing local spatial patterns in images?

ann.LSTM

bnn.Embedding

cnn.Conv2d

dnn.LayerNorm

Deep LearningEasy

Why is `optimizer.zero_grad()` called at the start of each training step?

aBecause PyTorch accumulates gradients across backward calls by default

bTo reset the learning rate

cTo re-initialise the weights randomly

dTo free the model from GPU memory

Deep LearningEasy

What does `model.eval()` change?

aIt disables gradient computation entirely

bIt freezes all parameters so they cannot be trained

cIt moves the model to the CPU

dIt switches layers like dropout and batch norm to inference behaviour

Deep LearningMedium

What does this code print?

import torch
x = torch.tensor(3.0, requires_grad=True)
y = x ** 2 + 4 * x
y.backward()
print(x.grad)

atensor(21.)

btensor(10.)

ctensor(6.)

dNone

Deep LearningMedium

What is the output shape?

import torch, torch.nn as nn
conv = nn.Conv2d(3, 16, kernel_size=3, padding=1)
x = torch.randn(8, 3, 64, 64)
print(conv(x).shape)

atorch.Size([8, 16, 64, 64])

btorch.Size([8, 16, 62, 62])

ctorch.Size([8, 3, 64, 64])

dtorch.Size([16, 8, 64, 64])

Deep LearningMedium

What inputs does `nn.CrossEntropyLoss` expect for multi-class classification?

aSoftmax probabilities and one-hot float labels

bSigmoid outputs and float labels

cRaw logits and integer class indices

dLog-probabilities and string labels

Deep LearningMedium

You have 400 labelled images across 4 classes. What is the best first approach?

aTrain a deep CNN from scratch for 200 epochs

bUse an RNN over the pixel rows

cTrain a transformer from scratch

dUse a pretrained backbone, freeze it, and train a new classification head

Deep LearningHard

Why are attention scores divided by the square root of the key dimension?

aTo stop large dot products from pushing softmax into saturated regions with tiny gradients

bTo make attention weights sum to 1

cTo reduce memory usage from quadratic to linear

dTo add positional information to the tokens

Deep LearningHard

What shapes are printed?

import torch, torch.nn as nn
lstm = nn.LSTM(input_size=10, hidden_size=32, num_layers=2, batch_first=True)
x = torch.randn(4, 15, 10)
out, (h, c) = lstm(x)
print(out.shape, h.shape)

a(4, 15, 32) and (4, 2, 32)

b(15, 4, 32) and (2, 4, 32)

c(4, 15, 32) and (2, 4, 32)

d(4, 32) and (2, 32)