Module: tf.contrib.mixed_precision
Stay organized with collections
Save and categorize content based on your preferences.
Library for mixed precision training.
Classes
class ExponentialUpdateLossScaleManager: Loss scale manager uses an exponential update strategy.
class FixedLossScaleManager: Loss scale manager with a fixed loss scale.
class LossScaleManager: Abstract loss scale manager class.
class LossScaleOptimizer: An optimizer that applies loss scaling in backprop.
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 2020-10-01 UTC.
[null,null,["Last updated 2020-10-01 UTC."],[],[],null,["# Module: tf.contrib.mixed_precision\n\n\u003cbr /\u003e\n\n|-------------------------------------------------------------------------------------------------------------------------------|\n| [View source on GitHub](https://github.com/tensorflow/tensorflow/blob/v1.15.0/tensorflow/contrib/mixed_precision/__init__.py) |\n\nLibrary for mixed precision training.\n\nClasses\n-------\n\n[`class ExponentialUpdateLossScaleManager`](../../tf/contrib/mixed_precision/ExponentialUpdateLossScaleManager): Loss scale manager uses an exponential update strategy.\n\n[`class FixedLossScaleManager`](../../tf/contrib/mixed_precision/FixedLossScaleManager): Loss scale manager with a fixed loss scale.\n\n[`class LossScaleManager`](../../tf/contrib/mixed_precision/LossScaleManager): Abstract loss scale manager class.\n\n[`class LossScaleOptimizer`](../../tf/contrib/mixed_precision/LossScaleOptimizer): An optimizer that applies loss scaling in backprop."]]
View source on GitHub