TensorFlow Image Classification Quiz

Submit

2. Which technique involves freezing weights of earlier layers when performing transfer learning?

Fine-tuning with frozen base layers

Data augmentation

Batch normalization

Learning rate scheduling

Submit

4. What is the kernel size in a Conv2D layer that captures fine-grained details?

1x1 kernel

3x3 kernel

7x7 kernel

11x11 kernel

5. True or False: Using a very large learning rate typically leads to better convergence in image classification models.

True

False

6. Which TensorFlow API allows you to load pre-trained models for transfer learning?

Tf.keras.applications

Tf.keras.layers

Tf.keras.optimizers

Tf.keras.callbacks

7. What is the purpose of flattening a layer before connecting to dense layers?

To convert 3D feature maps into 1D vectors for fully connected layers

To reduce the total number of parameters

To improve activation function performance

To eliminate the need for convolution

Submit

9. Which metric is most appropriate for evaluating imbalanced image classification datasets?

Accuracy only

Precision, recall, and F1-score

Loss value alone

Training speed

10. What does batch normalization accomplish in a CNN?

Reduces internal covariate shift and accelerates training

Increases the receptive field of convolutional filters

Eliminates the need for activation functions

Automatically adjusts learning rate during training

11. What is the primary advantage of using convolutional layers in image classification models?

They reduce spatial dimensions while preserving features

They capture local patterns through parameter sharing

They eliminate the need for activation functions

They automatically label images without training

12. What is the role of dropout in preventing overfitting?

It randomly deactivates neurons during training to reduce co-adaptation

It removes low-importance feature maps permanently

It scales all weights by a fixed factor

It applies convolution with random kernels

13. Which loss function is appropriate for multi-class image classification with mutually exclusive classes?

Binary crossentropy

Categorical crossentropy

Mean squared error

Hinge loss

14. What does the softmax activation function do in the output layer of a multi-class classifier?

Converts logits to probabilities that sum to 1

Applies a threshold to create binary predictions

Reduces overfitting by adding noise

Normalizes input features to zero mean

15. Which pre-trained model is commonly used for transfer learning in image classification?

VGG16

BERT

GPT-2

Word2Vec

Submit

17. What is the purpose of data augmentation in image classification?

To reduce training time significantly

To increase dataset size and improve model generalization

To automatically label unlabeled images

To eliminate the need for validation data

18. Which TensorFlow/Keras layer applies a learnable bias and activation function to feature maps?

Conv2D

Dense

Dropout

BatchNormalization

19. What does a max pooling layer do in a CNN?

Increases the number of feature maps

Selects the maximum value from each window to reduce dimensions

Applies convolution with larger kernels

Normalizes activations across the batch

20. In TensorFlow, which function normalizes pixel values to the range [0, 1]?

Tf.keras.preprocessing.image.normalize()

Image / 255.0

Tf.nn.batch_normalization()

Tf.image.per_image_standardization()

TensorFlow Developer Image Classification Quiz

1. In TensorFlow, the ImageDataGenerator class is used to ____ images during training.

2.

What first name or nickname would you like us to use?

2. Which technique involves freezing weights of earlier layers when performing transfer learning?

3. In TensorFlow, early stopping monitors validation ____ to prevent overfitting.

4. What is the kernel size in a Conv2D layer that captures fine-grained details?

5. True or False: Using a very large learning rate typically leads to better convergence in image classification models.

6. Which TensorFlow API allows you to load pre-trained models for transfer learning?

7. What is the purpose of flattening a layer before connecting to dense layers?

8. In TensorFlow, model.fit() trains the model using and to measure performance.

9. Which metric is most appropriate for evaluating imbalanced image classification datasets?

10. What does batch normalization accomplish in a CNN?

11. What is the primary advantage of using convolutional layers in image classification models?

12. What is the role of dropout in preventing overfitting?

13. Which loss function is appropriate for multi-class image classification with mutually exclusive classes?

14. What does the softmax activation function do in the output layer of a multi-class classifier?

15. Which pre-trained model is commonly used for transfer learning in image classification?

16. In TensorFlow, tf.data.Dataset.AUTOTUNE is used to optimize ____.

17. What is the purpose of data augmentation in image classification?

18. Which TensorFlow/Keras layer applies a learnable bias and activation function to feature maps?

19. What does a max pooling layer do in a CNN?

20. In TensorFlow, which function normalizes pixel values to the range [0, 1]?

TensorFlow Developer Image Classification Quiz

1. In TensorFlow, the ImageDataGenerator class is used to ____ images during training.

2.

What first name or nickname would you like us to use?

2. Which technique involves freezing weights of earlier layers when performing transfer learning?

3. In TensorFlow, early stopping monitors validation ____ to prevent overfitting.

4. What is the kernel size in a Conv2D layer that captures fine-grained details?

5. True or False: Using a very large learning rate typically leads to better convergence in image classification models.

6. Which TensorFlow API allows you to load pre-trained models for transfer learning?

7. What is the purpose of flattening a layer before connecting to dense layers?

8. In TensorFlow, model.fit() trains the model using ____ and ____ to measure performance.

9. Which metric is most appropriate for evaluating imbalanced image classification datasets?

10. What does batch normalization accomplish in a CNN?

11. What is the primary advantage of using convolutional layers in image classification models?

12. What is the role of dropout in preventing overfitting?

13. Which loss function is appropriate for multi-class image classification with mutually exclusive classes?

14. What does the softmax activation function do in the output layer of a multi-class classifier?

15. Which pre-trained model is commonly used for transfer learning in image classification?

16. In TensorFlow, tf.data.Dataset.AUTOTUNE is used to optimize ____.

17. What is the purpose of data augmentation in image classification?

18. Which TensorFlow/Keras layer applies a learnable bias and activation function to feature maps?

19. What does a max pooling layer do in a CNN?

20. In TensorFlow, which function normalizes pixel values to the range [0, 1]?

8. In TensorFlow, model.fit() trains the model using and to measure performance.