setu.discrete_nle_to_numpyro#
- setu.discrete_nle_to_numpyro(discrete_nle, theta, x_observed, *, conditions=None, name='discrete_nle_likelihood')[source]#
Add DiscreteNLE log-likelihood as a factor in the active NumPyro model.
Delegates to to_numpyro (which accepts any likelihood estimator). DiscreteNLE shares the log_prob interface with NLE/MixedNLE.
- Parameters:
discrete_nle (
LikelihoodEstimator) – Trained DiscreteNLE instance.theta (
Array) – JAX array from numpyro.sample(), shape (theta_dim,).x_observed (
Array) – Observed discrete data, shape (n_obs, x_dim) or (x_dim,).conditions (
Array|None) – Optional conditions, shape (n_obs, condition_dim).name (
str) – Name for the factor site.
- Return type: