tf.compat.v1.flags.FlagHolder | TensorFlow v2.16.1 (original) (raw)

tf.compat.v1.flags.FlagHolder

Stay organized with collections Save and categorize content based on your preferences.

Holds a defined flag.

tf.compat.v1.flags.FlagHolder(
    flag_values: tf.compat.v1.flags.FlagValues,
    flag: Flag[_T],
    ensure_non_none_value: bool = False
)

This facilitates a cleaner api around global state. Instead of::

flags.DEFINE_integer('foo', ...)
flags.DEFINE_integer('bar', ...)

def method():
  # prints parsed value of 'bar' flag
  print(flags.FLAGS.foo)
  # runtime error due to typo or possibly bad coding style.
  print(flags.FLAGS.baz)

it encourages code like::

_FOO_FLAG = flags.DEFINE_integer('foo', ...)
_BAR_FLAG = flags.DEFINE_integer('bar', ...)

def method():
  print(_FOO_FLAG.value)
  print(_BAR_FLAG.value)

since the name of the flag appears only once in the source code.

Args
flag_values	The container the flag is registered to.
flag	The flag object for this flag.
ensure_non_none_value	Is the value of the flag allowed to be None.

Attributes
default	Returns the default value of the flag.
name
present	Returns True if the flag was parsed from command-line flags.
value	Returns the value of the flag.If _ensure_non_none_value is True, then return value is notNone.

Methods

`serialize`

serialize() -> Text

Returns a serialized representation of the flag.

`bool`

__bool__()

`eq`

__eq__(
    other
)

Return self==value.

`nonzero`

__nonzero__()

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 2024-04-26 UTC.