Get the latest tech news
Numpyro: Probabilistic programming with NumPy powered by Jax
Probabilistic programming with NumPy powered by JAX for autograd and JIT compilation to GPU/TPU/CPU. - pyro-ppl/numpyro
One of the motivations for NumPyro was to speed up Hamiltonian Monte Carlo by JIT compiling the verlet integrator that includes multiple gradient computations. Note that for the class of distributions with loc,scale parameters such as Normal, Cauchy, StudentT, we also provide a LocScaleReparam reparameterizer to achieve the same purpose. The motivating ideas behind NumPyro and a description of Iterative NUTS can be found in this paper that appeared in NeurIPS 2019 Program Transformations for Machine Learning Workshop.
Or read this on Hacker News