Convolution layer (CONV) The convolution layer (CONV) works by using filters that perform convolution operations as it truly is scanning the enter $I$ with respect to its dimensions. Its hyperparameters contain the filter size $File$ and stride $S$. The ensuing output $O$ is called function map or activation map. https://financefeeds.com/carv-announces-the-launch-of-carv-svm-chain-testnet-empowering-ai-agents-redefining-data-sovereignty/