TensorFlow Questions et reponses d'entretien
Differences associees
Question 11. Explain the difference between eager execution and graph execution in TensorFlow.
Eager execution evaluates operations immediately, while graph execution involves building a computational graph to be executed later. Eager execution is the default mode in TensorFlow 2.x.
Example:
Question 12. What is the purpose of the tf.GradientTape in TensorFlow?
tf.GradientTape is used for automatic differentiation in TensorFlow. It records operations for computing gradients, allowing you to calculate gradients with respect to variables.
Example:
with tf.GradientTape() as tape:
tape.watch(x)
y = x * x
dy_dx = tape.gradient(y, x)
Question 13. Explain the concept of model checkpoints in TensorFlow.
Model checkpoints are a way to save the current state of a model during training. They can be used to resume training, fine-tune a model, or deploy a trained model.
Example:
Question 14. What is the purpose of the tf.function decorator in TensorFlow 2.x?
The tf.function decorator is used to convert a Python function into a TensorFlow graph, allowing for better performance through graph execution and enabling graph optimizations.
Example:
def my_function(x):
return x * x
Question 15. Explain the concept of data augmentation in image classification using TensorFlow.
Data augmentation involves applying random transformations to input data during training to increase the diversity of the dataset. In image classification, this can include rotations, flips, and zooms.
Example:
tf.keras.layers.experimental.preprocessing.RandomFlip('horizontal'),
tf.keras.layers.experimental.preprocessing.RandomRotation(0.2),
])
Les plus utiles selon les utilisateurs :