The output of your convolutional layer is usually handed in the ReLU activation purpose to bring non-linearity to the model. It's going to take the characteristic map and replaces many of the detrimental values with zero. It was observed that with the network depth expanding, the accuracy receives saturated https://financefeeds.com/new-meme-coin-copyright-all-stars-rallies-26-after-cex-listing-announcement/