Practice
This document has questions.
,Question-1
Statement
An image is a collection of pixels. A pixel is stored as a float value and typically occupies 4 bytes of
memory. Consider a dataset of images, where each image has dimensions .
Approximately, how much memory does the entire dataset occupy?
Options
(a)
4 KB
(b)
4 MB
(c)
40 MB
(d)
4 GB
Answer
(c)
Solution
We require float values to represent one image. Since each float value
occupies bytes of memory, a single image occupies bytes of memory. Roughly, this
corresponds to KB. The entire dataset would occupy KB or MB of memory. Here,
we have used the following facts:
KB bytes
MB KB
, Question-2
Statement
Consider a dataset that has points that belong to . All of them are found to lie on a line
that passes through the origin. We use a unit vector along the line as a representative and the
coefficients with respect to it to represent the individual data-points. Compute the percentage
decrease in the size of the dataset if we move to this new representation. Assume that it takes
one unit of space to store one feature. Enter your answer correct to two decimal places; it should
be in the range .
Answer
Range:
Solution
The size of the dataset in its original form is:
The size of the dataset after moving to the new representation:
The percentage decrease in the size of the dataset is therefore:
This document has questions.
,Question-1
Statement
An image is a collection of pixels. A pixel is stored as a float value and typically occupies 4 bytes of
memory. Consider a dataset of images, where each image has dimensions .
Approximately, how much memory does the entire dataset occupy?
Options
(a)
4 KB
(b)
4 MB
(c)
40 MB
(d)
4 GB
Answer
(c)
Solution
We require float values to represent one image. Since each float value
occupies bytes of memory, a single image occupies bytes of memory. Roughly, this
corresponds to KB. The entire dataset would occupy KB or MB of memory. Here,
we have used the following facts:
KB bytes
MB KB
, Question-2
Statement
Consider a dataset that has points that belong to . All of them are found to lie on a line
that passes through the origin. We use a unit vector along the line as a representative and the
coefficients with respect to it to represent the individual data-points. Compute the percentage
decrease in the size of the dataset if we move to this new representation. Assume that it takes
one unit of space to store one feature. Enter your answer correct to two decimal places; it should
be in the range .
Answer
Range:
Solution
The size of the dataset in its original form is:
The size of the dataset after moving to the new representation:
The percentage decrease in the size of the dataset is therefore:
