Question 25 (Advanced - CNN Implementation in Keras):
When building a CNN for image classification in Keras, what is the purpose of Global Average Pooling 2D as the final layer before classification?

A) Reduces spatial dimensions to 1x1 while preserving channel depth
B) Increases receptive field for better feature extraction
C) Performs pixel-wise normalization
D) Adds non-linearity before dense layers

#Python #Keras #CNN #DeepLearning

✅ By: https://yangx.top/DataScienceQ