Why Keras simplifies model building in TensorFlow - Challenge Your Understanding

Challenge - 5 Problems

🧠 Conceptual

intermediate

Why does Keras use a Sequential model?

Keras provides a Sequential model to build neural networks. What is the main reason this model simplifies building layers?

A It only supports convolutional layers, limiting complexity.

B It allows stacking layers one after another in a simple linear way without managing connections manually.

C It requires writing low-level code for each neuron connection.

D It automatically tunes hyperparameters without user input.

💡 Hint

Think about how layers are added in a simple list.

The Sequential model lets you add layers in order, so you don't have to manually connect each layer's inputs and outputs.

❓ Predict Output

intermediate

Output shape after adding layers in Keras Sequential

What will be the output shape of the model after running this code?

TensorFlow

from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
model = Sequential()
model.add(Dense(10, input_shape=(5,)))
model.add(Dense(3))
output_shape = model.output_shape

A (3, 5)

B (5, 10)

C (None, 10)

D (None, 3)

💡 Hint

The output shape shows batch size as None and last layer units.

The last Dense layer has 3 units, so output shape is (None, 3), where None is batch size.

❓ Model Choice

advanced

Choosing Keras Functional API for complex models

Why would you choose Keras Functional API over Sequential model?

A To build models with multiple inputs or outputs and non-linear layer connections.

B To simplify stacking layers in a straight line.

C To automatically generate training data.

D To avoid writing any code for model training.

💡 Hint

Think about models that are not just a simple chain of layers.

Functional API allows building complex architectures like branching or merging layers, which Sequential cannot do.

❓ Hyperparameter

advanced

Effect of batch size in Keras model training

What is the main effect of increasing batch size during Keras model training?

A It changes the activation function of layers.

B It increases the model's number of layers automatically.

C It reduces the number of weight updates per epoch, potentially speeding up training but may reduce generalization.

D It disables dropout layers during training.

💡 Hint

Think about how many samples are processed before updating weights.

Larger batch sizes mean fewer updates per epoch, which can speed training but might hurt how well the model generalizes.

❓ Metrics

expert

Interpreting Keras model accuracy during training

During training, a Keras model shows training accuracy increasing but validation accuracy stays flat or decreases. What does this indicate?

A The model is overfitting the training data and not generalizing well to new data.

B The model is underfitting and needs more layers.

C The optimizer is not updating weights correctly.

D The batch size is too small causing unstable training.

💡 Hint

Think about what it means when training improves but validation does not.

When training accuracy improves but validation accuracy does not, the model memorizes training data but fails on unseen data, a sign of overfitting.