Faster Hamiltonian Monte Carlo by Learning Leapfrog Scale

Hamiltonian Monte Carlo samplers have become standard algorithms for MCMC implementations, as opposed to more bas