This is a fully parameterized verilog implementation of computation kernels for accleration of the Inference of Convolutional Neural Networks on FPGAs