Learning Bayes' theorem with a neural network for gravitational-wave inference