tf.data.Options

Class Options

Defined in tensorflow/python/data/ops/dataset_ops.py.

Represents options for tf.data.Dataset.

An Options object can be for instance used to control which static optimizations to apply or whether to use performance modeling to dynamically tune the parallelism of operations such as tf.data.Dataset.map or tf.data.Dataset.interleave.

__init__

__init__()

Initialize self. See help(type(self)) for accurate signature.

Properties

experimental_autotune

Whether to dynamically adjust the values of tunable parameters (e.g. degrees of parallelism).

experimental_filter_fusion

Whether to fuse filter transformations.

experimental_hoist_random_uniform

Whether to hoist tf.random_uniform() ops out of map transformations.

experimental_latency_all_edges

Whether to add latency measurements on all edges.

experimental_map_and_batch_fusion

Whether to fuse map and batch transformations.

experimental_map_and_filter_fusion

Whether to fuse map and filter transformations.

experimental_map_fusion

Whether to fuse map transformations.

experimental_map_parallelization

Whether to parallelize stateless map transformations.

experimental_map_vectorization

Whether to vectorize map transformations.

experimental_noop_elimination

Whether to eliminate no-op transformations.

experimental_shuffle_and_repeat_fusion

Whether to fuse shuffle and repeat transformations.

Methods

__eq__

__eq__(other)

Return self==value.

__ne__

__ne__(other)

Return self!=value.

merge

merge(options)

Merges itself with the given tf.data.Options.

The given tf.data.Options can be merged as long as there does not exist an attribute that is set to different values in self and options.

Args:

Raises:

Returns:

New tf.data.Options() object which is the result of merging self with the input tf.data.Options.