You're building a CNN to classify handwritten digits (MNIST). You extract activations from two convolutional layers:
What difference do you expect between activation_1 and activation_2?

Question

You're building a CNN to classify handwritten digits (MNIST). You extract activations from two convolutional layers: What difference do you expect between activation_1 and activation_2?

Answer

They should be identical since both are convolutional

Answer

activation_2 captures higher-level features like corners or digits

Answer

activation_1 has more channels than activation_2

Answer

activation_1 contains more semantic information

Crowdly

You're building a CNN to classify handwritten digits (MNIST). You extract activ...

Want instant access to all verified answers on moodle.astanait.edu.kz?