Feature map visualization in TensorFlow - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the TensorFlow library.

TensorFlow

import [1] as tf

Drag options to blanks, or click blank then click option'

Akeras

Btorch

Cnumpy

Dtensorflow

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'torch' instead of 'tensorflow' will import PyTorch, not TensorFlow.

Using 'numpy' or 'keras' alone won't give access to TensorFlow functions.

We use import tensorflow as tf to access TensorFlow functions easily.

2fill in blank

medium

Complete the code to create a simple convolutional layer named 'conv_layer'.

TensorFlow

conv_layer = tf.keras.layers.Conv2D(filters=32, kernel_size=3, activation=[1])

Drag options to blanks, or click blank then click option'

A'relu'

B'sigmoid'

C'softmax'

D'tanh'

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'softmax' activation in Conv2D is uncommon and usually for output layers.

Choosing 'sigmoid' or 'tanh' may slow down training for convolutional layers.

The 'relu' activation is commonly used in convolutional layers for feature extraction.

3fill in blank

hard

Fix the error in the code to get the output feature maps from the convolutional layer.

TensorFlow

feature_maps = conv_layer([1])

Drag options to blanks, or click blank then click option'

Ainput_image.numpy()

Binput_image.shape

Cinput_image

Dinput_image.data

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the shape or numpy array causes errors or unexpected behavior.

Using '.data' attribute is not valid for TensorFlow tensors.

We pass the actual input tensor input_image to the layer to get feature maps.

4fill in blank

hard

Fill both blanks to plot the first 6 feature maps using matplotlib.

TensorFlow

import matplotlib.pyplot as plt
plt.figure(figsize=(12, 6))
for i in range([1]):
    plt.subplot(1, 6, i+1)
    plt.imshow(feature_maps[0, :, :, i], cmap=[2])
    plt.axis('off')
plt.show()

Drag options to blanks, or click blank then click option'

A6

B'gray'

C'viridis'

D3

Attempts:

3 left

💡 Hint

Common Mistakes

Using a range less than 6 will miss some feature maps.

Using color maps like 'viridis' may make features less clear.

We plot 6 feature maps and use 'gray' colormap for clear visualization.

5fill in blank

hard

Fill all three blanks to extract feature maps from a model layer named 'conv2d_1' for input 'img_tensor'.

TensorFlow

from tensorflow.keras.models import Model
layer_output = model.get_layer([1]).output
activation_model = Model(inputs=model.input, outputs=[2])
feature_maps = activation_model.predict([3])

Drag options to blanks, or click blank then click option'

A'conv2d_1'

Blayer_output

Cimg_tensor

D'conv2d_2'

Attempts:

3 left

💡 Hint

Common Mistakes

Using wrong layer names causes errors.

Passing wrong inputs to predict causes runtime errors.

We get the output of 'conv2d_1' layer, create a model for activations, and predict on 'img_tensor'.