29. Pooling Layers

Other kinds of pooling

Alexis mentioned one other type of pooling, and it is worth noting that some architectures choose to use average pooling , which chooses to average pixel values in a given window size. So in a 2x2 window, this operation will see 4 pixel values, and return a single, average of those four values, as output!

This kind of pooling is typically not used for image classification problems because maxpooling is better at noticing the most important details about edges and other features in an image, but you may see this used in applications for which smoothing an image is preferable.

Notebook: Layer Visualizations

The next notebooks are about visualizing the output of convolutional and pooling layers.

To open the notebook, you have two options:

• Go to the next page in the classroom (recommended).
• Clone the repo from Github and open the notebook conv_visualization.ipynb & maxpooling_visualization.ipynb in the convolutional-neural-networks > conv-visualization folder. You can either download the repository with git clone https://github.com/udacity/deep-learning-v2-pytorch.git , or download it as an archive file from this link .

Instructions

This notebook is meant to give you a chance to explore the effect of convolutional layers, activations, and pooling layers!