Given two continous probability distributions $p(x)$ and $q(x)$ their KL divergence is $$KL(p, q) = \int_x p(x) \log\left(\frac{p(x)}{q(x)}\right)$$