TensorFlow is back at Google I/O on May 14! Register now

Module: tfp.experimental.distributions.marginal_fns

Experimental functions to use as marginals for GaussianProcess(es).

Modules

mvn_linear_operator module: Multivariate Normal distribution classes.

ps module: Operations that use static values when possible.

tfp_custom_gradient module: TF and JAX compatible custom gradients.

make_backoff_cholesky(...): Make a function that tries Cholesky then the user-specified function.

make_cholesky_like_marginal_fn(...): Use a Cholesky-like function for GaussianProcess marginal_fn.

make_eigh_marginal_fn(...): Make an eigenvalue decomposition-based marginal_fn.

retrying_cholesky(...): Computes a modified Cholesky decomposition for a batch of square matrices.

Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

Last updated 2023-11-21 UTC.