Tensor math operations in TensorFlow - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

What is the output of this TensorFlow code?

Consider the following code snippet using TensorFlow. What will be the printed output?

TensorFlow

import tensorflow as tf

a = tf.constant([[1, 2], [3, 4]])
b = tf.constant([[5, 6], [7, 8]])
c = tf.add(a, b)
print(c.numpy())

A

[[6 8]
 [10 12]]

B

[[4 8]
 [10 12]]

C

[[6 8]
 [11 12]]

D

[[5 7]
 [10 12]]

💡 Hint

tf.add adds corresponding elements of two tensors.

The tf.add function adds each element of tensor 'a' to the corresponding element of tensor 'b'. So 1+5=6, 2+6=8, 3+7=10, 4+8=12.

❓ Model Choice

intermediate

Which TensorFlow operation computes the element-wise product of two tensors?

You want to multiply two tensors element by element in TensorFlow. Which operation should you use?

A tf.matmul

B tf.multiply

C tf.add

D tf.reduce_sum

💡 Hint

Think about the operation that multiplies elements one by one.

tf.multiply performs element-wise multiplication, multiplying each element of one tensor by the corresponding element of the other tensor.

❓ Hyperparameter

advanced

Choosing the correct axis for tf.reduce_sum

Given a tensor of shape (3, 4, 5), which axis should you reduce over to get a tensor of shape (3, 4)?

A axis=0

B axis=1

C axis=None

D axis=2

💡 Hint

Reducing over an axis removes that dimension.

Reducing over axis=2 (the last dimension of size 5) sums across that dimension, resulting in shape (3, 4).

🔧 Debug

advanced

Why does this TensorFlow code raise an error?

Examine the code below. Why does it raise an error?

TensorFlow

import tensorflow as tf

a = tf.constant([1, 2, 3])
b = tf.constant([[1, 2], [3, 4], [5, 6]])
c = tf.add(a, b)
print(c.numpy())

A 'a' and 'b' have different data types

B tf.add cannot add tensors with more than one dimension

C Shapes of 'a' and 'b' are incompatible for broadcasting

D TensorFlow requires explicit reshaping before addition

💡 Hint

Check the shapes of the tensors and how broadcasting works.

Tensor 'a' has shape (3,) and 'b' has shape (3,2). TensorFlow tries to broadcast but the shapes are incompatible because the first dimension of 'a' is missing.

🧠 Conceptual

expert

What is the effect of tf.transpose on a 3D tensor?

You have a tensor of shape (2, 3, 4). What will be the shape after applying tf.transpose with perm=[1, 0, 2]?

A (3, 2, 4)

B (4, 3, 2)

C (2, 4, 3)

D (3, 4, 2)

💡 Hint

tf.transpose permutes the dimensions according to the perm list.

The perm argument [1, 0, 2] swaps the first and second dimensions, so shape (2,3,4) becomes (3,2,4).