I am currently training a neural network that has to output a binary classification. When the model is trained the binary_accuracy that is built in keras converges to the la
