What does CONVOLUTION_NUMBER indicate while visualising the layers from the model?

Course: Introduction to TensorFlow for Artificial Intelligence, Machine Learning, and Deep Learning, Week 3, Lab 1

@nsrinidhibhat Welcome to the community
Please be more specific and mention the course and week details for your query.

1 Like

Thank you. I have edited my question to provide for the course and week.

CONVOLUTION_NUMBER represents the index of the feature map of the convolution layer that you want to visualize.

1 Like

Could you explain more? I don’t understand what you mean by “feature map of the convolution layer.” Since the image is 28x28, I assumed CONVOLUTION_NUMBER = 30 would be out of bounds. It was not out of bounds. Then I tried CONVOLUTION_NUMBER = 100 and it became out of bounds. I don’t understand the relationship between CN and the images.

In convolutional neural networks, you have filters/kernels of different shapes [(1, 1), (3, 3), (5, 5), etc.] that convolve with the image to extract features. The extracted features are called feature maps. You can have many such filters in each layer and each filter will generate its own feature map.

In your case, the convolution layer has 64 filters, if you choose CONVOLUTION_NUMBER beyond 63, it will become out of bound.

1 Like