Questions and published explanations below are available without starting a test. Some questions may not have a published solution yet.
0 / 22 published solutions
Question 2 MCQ · 5.0 marks
In the context of Single Image Super-Resolution (SISR), specifically in the ”Pixel Shuffle” layer (used
in models like SRResNet), how are the tensor dimensions transformed?
6406536039865.
From (H, W, C) to (H/r, W/r,
)
6406536039866.
From (H, W,
) to (H · r, W · r, C)
6406536039867.
From (H, W, C) to (H · r, W · r, C) using bilinear interpolation
6406536039868.
From (H, W, C) to (H, W, C/r) using max pooling
A published solution is not available for this question yet.
Which of the following equations correctly represents the building block of a Residual Network
(ResNet), where x is input and F(x) is the residual mapping?
6406536039873.
H(x) = F(x)
6406536039874.
H(x) = F(x) + x
6406536039875.
H(x) = F(x) · x
6406536039876.
H(x) = F(x) - x
A published solution is not available for this question yet.
AlexNet introduced ”Local Response Normalization (LRN)”. Although less common now (often
replaced by Batch Norm), what was its original intended purpose?
6406536039881.
To aid generalization by creating competition for big activities amongst
neuron outputs computed using different kernels.
6406536039882.
To prevent the vanishing gradient problem in deep layers.
6406536039883.
To reduce the number of parameters in the fully connected layers.
6406536039884.
To strictly enforce a mean of 0 and variance of 1 across the batch.
A published solution is not available for this question yet.
Consider an input image of size 227 × 227. You apply a convolution with a kernel size of 11 × 11, a
stride of 4, and padding of 0. What is the spatial width/height of the output feature map?
A published solution is not available for this question yet.
In the original YOLO architecture, the image is split into a 7 × 7 grid. If the model is configured to
predict B = 2 bounding boxes per grid cell, what is the total number of bounding boxes predicted
for the entire image?
A published solution is not available for this question yet.
A feature map has dimensions 14 × 14 × 512. We apply a 1 × 1 convolution layer with 128 filters.
How many parameters (weights only, ignore bias) does this layer have?
A published solution is not available for this question yet.
Read the scenario below and answer the given subquestions.
A ”Cat vs Dog” classifier is tested on a dataset containing 100 images (80 Cats, 20 Dogs). The model
predicts ”Cat” for 85 images and ”Dog” for 15 images. Out of the 85 ”Cat” predictions, 75 are
actually Cats.
Calculate the Precision for the ”Cat” class. (Round to 2 decimal places).
A published solution is not available for this question yet.
Read the scenario below and answer the given subquestions.
A ”Cat vs Dog” classifier is tested on a dataset containing 100 images (80 Cats, 20 Dogs). The model
predicts ”Cat” for 85 images and ”Dog” for 15 images. Out of the 85 ”Cat” predictions, 75 are
actually Cats.
Calculate the Recall for the ”Cat” class. (Round to 2 decimal places).
A published solution is not available for this question yet.
Read the scenario below and answer the given subquestions.
A ”Cat vs Dog” classifier is tested on a dataset containing 100 images (80 Cats, 20 Dogs). The model
predicts ”Cat” for 85 images and ”Dog” for 15 images. Out of the 85 ”Cat” predictions, 75 are
actually Cats.
Calculate the F1 Score for the ”Cat” class. (Round to 2 decimal places).
A published solution is not available for this question yet.
