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

tf.compat.v1.flags.flag_dict_to_args

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

Convert a dict of values into process call parameters.

tf.compat.v1.flags.flag_dict_to_args(
    flag_map: Dict[str, Any], multi_flags: Optional[Set[str]] = None
) -> Iterable[str]

This method is used to convert a dictionary into a sequence of parameters for a binary that parses arguments using this module.

Args
flag_map	dict, a mapping where the keys are flag names (strings). values are treated according to their type: If value is None, then only the name is emitted. If value is True, then only the name is emitted. If value is False, then only the name prepended with 'no' is emitted. If value is a string then --name=value is emitted. If value is a collection, this will emit--name=value1,value2,value3, unless the flag name is inmulti_flags, in which case this will emit--name=value1 --name=value2 --name=value3. Everything else is converted to string an passed as such.
multi_flags	set, names (strings) of flags that should be treated as multi-flags.

Args

flag_map

dict, a mapping where the keys are flag names (strings). values are treated according to their type: If value is None, then only the name is emitted. If value is True, then only the name is emitted. If value is False, then only the name prepended with 'no' is emitted. If value is a string then --name=value is emitted. If value is a collection, this will emit--name=value1,value2,value3, unless the flag name is inmulti_flags, in which case this will emit--name=value1 --name=value2 --name=value3. Everything else is converted to string an passed as such.

multi_flags

set, names (strings) of flags that should be treated as multi-flags.

Yields
sequence of string suitable for a subprocess execution.

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.