Skip to content

GitLab

Commit a76b94d7 authored by Yang Chen's avatar Yang Chen Committed by A. Unique TensorFlower
Browse files

Internal change 

PiperOrigin-RevId: 474325161
parent fc71d88d
Show whitespace changes
Inline Side-by-side
Showing with 1 addition and 1 deletion
official/core/config_definitions.py
View file @ a76b94d7
...@@ -89,7 +89,7 @@ class DataConfig(base_config.Config): ...@@ -89,7 +89,7 @@ class DataConfig(base_config.Config):
tf.data pipeline. When using shared tf.data service, the tf.data dataset tf.data pipeline. When using shared tf.data service, the tf.data dataset
must be infinite, and slow trainer may skip certain training examples. must be infinite, and slow trainer may skip certain training examples.
More details about shared tf.data service can be found at: More details about shared tf.data service can be found at:
https://www.tensorflow.org/api_docs/python/tf/data/experimental/service?version=nightly#sharing_tfdata_service_with_concurrent_trainers. https://www.tensorflow.org/api_docs/python/tf/data/experimental/service#sharing_tfdata_service_with_concurrent_trainers.
apply_tf_data_service_before_batching: A bool. If set to True, tf.data apply_tf_data_service_before_batching: A bool. If set to True, tf.data
service will be applied before batching operation. This is useful to make service will be applied before batching operation. This is useful to make
sure only a single tf.data service instance is created when sure only a single tf.data service instance is created when
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment