Module: tfp.experimental.auto_batching.numpy_backend

Numpy backend for auto-batching VM.

It can be faster than TF for tiny examples and prototyping, and moderately simpler due to immediate as opposed to deferred result computation.

All operations take and ignore name= arguments to allow for useful op names in the TensorFlow backend.

Classes

class NumpyBackend: Implements the Numpy backend ops for a PC auto-batching VM.

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.