- Sample posteriors with SVI in numpyro
- In this example, we have an observed stellar spectrum that we would like to fit as a mixture of models for one cool and one hot stellar spectrum. We have two PHOENIX model stellar spectra which bracket the solar effective temperature, and a publicly available "observed" solar spectrum.
- How to compute leave-one-out cross validation stats for a numpyro (jax) model with a Gaussian process from celerite2
- Based on the Leave-One-Out article in the stan docs
- Pareto smoothed importance sampling leave-one-out cross validation
- In this tutorial, we will construct a linear model with numpyro and carry out Pareto-Smoothed importance sampling Leave-One-Out Cross Validation on the results.
- These slides were part of the presentation of this notebook
- Introduction to JAX for physical scientists
- Interactive MCMC demo
- Interactive Gaussian process demo
- Interactive Approximate Bayesian Computation demo
- Python for Astronomy Tutorial at Exeter (2019)
- Intro to Astropy Affiliated Packages at .Astronomy 9
- Linear model selection with PyMC3