tf.experimental.dtensor.enable_save_as_bf16
Stay organized with collections
Save and categorize content based on your preferences.
Allows float32 DVariables to be checkpointed and restored as bfloat16.
tf.experimental.dtensor.enable_save_as_bf16(
variables: List[tf.Variable]
)
The method only affects the DVariable part inside the model and leaves
non-DTensor Variables/Tensors untouched.
Args |
|---|
variables
|
A list of tf.Variable to be enabled with bfloat16 save/restore.
Only has effect on DTensor Variables as they go through d_variables with
DTensor Specific logis.
|
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. Some content is licensed under the numpy license.
Last updated 2023-03-17 UTC.
[[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Missing the information I need","missingTheInformationINeed","thumb-down"],["Too complicated / too many steps","tooComplicatedTooManySteps","thumb-down"],["Out of date","outOfDate","thumb-down"],["Samples / code issue","samplesCodeIssue","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2023-03-17 UTC."],[],[]]
View source on GitHub