TensorFlow

API

 tf.compat / v1 / v1.flags.mark_flags_as_required


Ensures that flag is not None during program execution.

Registers a flag validator, which will follow usual validator rules. Important note: validator will pass for any non-None value, such as False, 0 (zero), '' (empty string) and so on.

If your module might be imported by others, and you only wish to make the flag required when the module is directly executed, call this method like this:

if name == 'main': flags.mark_flag_as_required('your_flag_name') app.run()

flag_name str, name of the flag
flag_values flags.FlagValues, optional FlagValues instance where the flag is defined.

AttributeError Raised when flag_name is not registered as a valid flag name.

此页内容是否对您有帮助