0
$\begingroup$

This is what Andrew Ng draws in his pooling layers video in the Coursera Deep Learning Specialization:

enter image description here

and this is what he draws in Inception network video:

enter image description here

Notice in first slide, number of input and output channels is same as pooling layers processes each channel independently and thus produces as many output channels as there are in the input. But in the second slide, the number of output and input channels of the MAX-POOL is different: number of input channels to MAX-POOL is 192 (encircled orange) and the number of output channels is 32 (encircled red). How in this example MAX-POOL outputted in different number of channels? What detail I am missing?

$\endgroup$
0
$\begingroup$

Without having seen the full video you are referring to I think the image represents the inception module with dimension reduction from the original paper (figure 2b from here, also shown below). enter image description here

Based on this it seems that not all convolutional layers have been shown in the image, mainly the 1x1 convolutional layer that follows the max-pooling layer (see also figure 4 from here). The max-pooling layer itself only decreases the height and the width of the incoming array and does not change the number of channels. The number of channels is simply reduced by the 1x1 convolutional layer with 32 filters that follows the max-pooling layer, which leads to the 32 channels you see in the image.

$\endgroup$
1
  • $\begingroup$ indeed... this is explained in the next video which is dedicated to more detailed discussion on inception network. Its very confusing to not to understand certain thing after watching the video several times to learn that its explained in the next video ... 😑 $\endgroup$
    – Rnj
    Jan 3 at 11:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.