esem.sampler.MCMCSampler.sample¶
- MCMCSampler.sample(prior_x=None, n_samples=1, kernel_kwargs=None, mcmc_kwargs=None)¶
This is the call that does the actual inference.
It should call model.sample over the prior, compare with the objective, and then output a posterior distribution
- Parameters
prior_x (
tensorflow_probability.distribution
) – The distribution to sample parameters from. By default it will uniformly sample the unit N-D hypercuben_samples (
int
) – The number of samples to drawkernel_kwargs (
dict
) – kwargs for the MCMC kernelmcmc_kwargs (
dict
) – kwargs for the MCMC sampler
- Returns
np.array
– Array of samples