VariationalGaussianInit#

class gpjax.variational_families.VariationalGaussianInit(inducing_inputs, variational_mean=None, variational_root_covariance=None)[source]#

Bases: object

Initialization parameters for a variational Gaussian distribution.

Expand for references to gpjax.variational_families.VariationalGaussianInit

Heteroscedastic Inference / Background

Parameters:

  • inducing_inputs (Int[jaxlib._jax.Array, 'N D'] | Int[ndarray, 'N D'] | Float[jaxlib._jax.Array, 'N D'] | Float[ndarray, 'N D'])

  • variational_mean (Float[jaxlib._jax.Array, 'N 1'] | Float[ndarray, 'N 1'] | None)

  • variational_root_covariance (Float[jaxlib._jax.Array, 'N N'] | Float[ndarray, 'N N'] | None)