To build a multilayer perceptron model and to train datas from it So first we define the architecture of the multi-layer Perceptron and then forward propogation is defined and also backward propogation The results are predicted for non-linear datasets and acuuracy is found