wayground logo

Free Printable Worksheets

Font size

S
M
L
XL
Worksheets

GLIM - PGPM - CV Quiz3_Transfer Learning models (DL)

Total questions: 11

Worksheet time: 6mins

Name
Class
Date
1.

What is the primary idea behind transfer learning in computer vision?

a)

Training a CNN from scratch on a small dataset

b)

Reusing knowledge from a model trained on a large dataset for a new task

c)

Increasing the number of hidden layers in a CNN

d)

Compressing a deep learning model for deployment

2.

Which dataset is most commonly used to pretrain models such as VGG, ResNet, and Inception?

a)

CIFAR-10

b)

MNIST

c)

ImageNet

d)

COCO

3.

Which part of a pretrained CNN is usually replaced for a new classification task?

a)

Convolution layers

b)

Pooling layers

c)

Final fully connected (classification) layer

d)

Input layer

4.

A major limitation of VGG compared to modern CNNs is:

a)

Poor accuracy

b)

Very small number of parameters

c)

High memory and computational requirements

d)

Inability to use transfer learning

5.

ResNet introduced which key concept to solve the vanishing gradient problem?

a)

Dropout

b)

Batch normalization

c)

Skip (residual) connections

d)

Max pooling

6.

YOLO is primarily designed for:

a)

Image classification only

b)

Semantic segmentation

c)

Real-time object detection

d)

Image compression

7.

A key difference between YOLO and CNN classifiers (VGG/ResNet) is that YOLO:

a)

Uses grayscale images only

b)

Performs detection and classification in a single forward pass

c)

Requires image patches as input

d)

Cannot be trained using transfer learning

8.

Which metric is most important when missing a critical class (e.g., large potholes or safety hazards)?

a)

Accuracy

b)

Precision

c)

Recall

d)

Loss value

9.

F1 score is preferred over accuracy when:

a)

Dataset is perfectly balanced

b)

Training time is high

c)

Model size is large

d)

Classes are imbalanced

10.

From a managerial perspective, the best model is the one that:

a)

Has the highest accuracy only

b)

Is the deepest network available

c)

Balances performance, speed, and deployment cost

d)

Uses the latest architecture

11.

Which model would be most suitable for deployment on a low-memory mobile device?

a)

MobileNet

b)

EfficientNet

c)

ResNet50

d)

VGG16