I was reading AlexNet paper and the authors quoted

the kernels on one GPU were "largely color agnostic," whereas the kernels on the other GPU were largely "color-specific."

The upper GPU takes operates on filters on the top and lower GPU deals with the lower half. But what is the reason for each of them learning a different set of features, i.e. the top half of kernels learning the edges mostly and the bottom kernels learning color variation? Is there any reason behind it?



Your Answer

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

Browse other questions tagged or ask your own question.