01In the Keras workflow, what happens during the 'Compile' step?
02What does an 'Epoch' represent in TensorFlow training?
03Which Callback would you use to save the best version of your model automatically?
04What is the primary benefit of using the 'tf.data' API?
05What is 'Transfer Learning'?
06Why is 'Normalization' (dividing pixel values by 255.0) crucial in image models?
07Which `tf.keras.datasets` would you use to practice with small color images of 10 different classes?
08What is the role of `layers.Flatten()` at the start of an image model?
09Which function is used to load a fully trained model from a file?
10In a multi-class classifier with 10 classes, what shape does `model.predict()` return for one image?