TensorFlow Advanced TensorFlow Techniques 3

Question 1

What is the purpose of tf.keras.utils.plot_model()?

Accepted Answer

Generates a visualization of the model architecture as an image

Answer

plot_model() renders a diagram of the model's layers and connections, saving it as a PNG or SVG file.

Question 2

What is the main advantage of using XLA (Accelerated Linear Algebra) with TensorFlow?

Accepted Answer

Compiles computation graphs into optimized machine code for faster execution

Answer

XLA JIT-compiles TF ops into optimized native code, fusing operations and reducing memory bandwidth for significant speedups on TPUs and GPUs.

Question 3

Which TensorFlow method computes gradients of a loss with respect to variables?

Accepted Answer

tape.gradient(loss, variables) inside tf.GradientTape

Answer

Inside a tf.GradientTape context, tape.gradient() computes derivatives of the recorded computation with respect to specified variables.

Question 4

What does tf.keras.applications provide?

Accepted Answer

Pre-built, pre-trained models like ResNet, VGG, and MobileNet for transfer learning

Answer

tf.keras.applications includes popular CNN architectures pre-trained on ImageNet that can be used directly or fine-tuned for custom tasks.

Question 5

What is the purpose of tf.keras.metrics.Mean()?

Accepted Answer

Tracks the running average of a scalar value across multiple update calls

Answer

tf.keras.metrics.Mean() accumulates values and computes their mean, useful for tracking average loss or custom metrics during training.

TensorFlow Practice Test