How to share a fully connected layer and is it possible to use a channel direction pooling?
Show older comments
(한국어로 답변주실 수 있다면 한국어로 답변 부탁드립니다.)
I am currently designing a model using the Deep Network Designer to implement a Convolutional Block Attention Module (CBAM).
For this, I need to use a channel-wise max/average pooling layer, and I am wondering if there is a way to implement this within MATLAB.
Additionally, I would like to know if there is a method to share a fully connected layer across different parts of the network. The current implementation of the fully connected layer seems to be limited to a single input and output.
Lastly, when passing through the fully connected layer, the dimension of the output is converted to (C x B). Is there a way to reshape this back to (S x S x C x B) in order to perform element-wise multiplication?
Accepted Answer
More Answers (0)
Categories
Find more on Deep Learning Toolbox in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
