TensorFlow & Keras

Q: Is the TensorFlow & Keras quiz free?

Yes. The TensorFlow & Keras quiz is completely free with no signup. Answer the questions, get instant scoring, and earn XP as you go.

Q: What level is the TensorFlow & Keras quiz?

This is a Medium-level quiz. Each question includes the correct answer so you learn as you test yourself, whatever your starting point.

Q: How can I prepare for this quiz?

Review the linked courses and notes for the topic first, then take the quiz to check your understanding. You can retake it anytime to track improvement.

Question 1

In Keras, what does model.compile() specify?

Accepted Answer

compile() configures how the model trains: which optimizer, which loss to minimize, and which metrics to report.

Question 2

For multi-class classification with integer labels, sparse_categorical_crossentropy avoids one-hot encoding.

Accepted Answer

True. sparse_categorical_crossentropy takes integer class labels directly, while categorical_crossentropy expects one-hot vectors.

Question 3

What is an epoch?

Accepted Answer

An epoch is one complete pass through all training samples; each epoch usually contains many mini-batch steps.

Question 4

Why normalize/scale input features before training?

Accepted Answer

Unscaled features distort the loss surface and slow/destabilize optimization; normalization speeds convergence.

Question 5

What does a softmax output layer produce for classification?

Accepted Answer

softmax converts logits into class probabilities that sum to 1, so the largest is the predicted class.

Question 6

tf.data pipelines can prefetch and batch data to keep the GPU fed.

Accepted Answer

True. tf.data with .batch().prefetch() overlaps preprocessing with training, avoiding input-pipeline bottlenecks.

Question 7

What is transfer learning?

Accepted Answer

You start from a model trained on a large dataset (e.g., ImageNet) and fine-tune it for your task — far less data/compute than training from scratch.

Question 8

What does an EarlyStopping callback do?

Accepted Answer

EarlyStopping halts training once validation performance plateaus, often restoring the best weights.

Question 9

A learning rate that is too high can make training diverge (loss goes to NaN).

Accepted Answer

True. Too-large steps overshoot minima and can blow up the loss; too-small learns slowly. Tuning the LR is crucial.

Question 10

What format is commonly used to deploy a trained TF model?

Accepted Answer

TensorFlow exports to SavedModel; TFLite targets mobile/edge and TF Serving hosts models for production inference.

TensorFlow & Keras

Frequently Asked Questions

Is the TensorFlow & Keras quiz free?

What level is the TensorFlow & Keras quiz?

How can I prepare for this quiz?

Learn the Topic

ML Basics Quiz

Go deeper on this topic

TensorFlow & Keras

Frequently Asked Questions

Is the TensorFlow & Keras quiz free?

What level is the TensorFlow & Keras quiz?

How can I prepare for this quiz?

Learn the Topic

ML Basics Quiz

Go deeper on this topic