pytorch_pfn_extras.training.triggers.FunctionTrigger#

class pytorch_pfn_extras.training.triggers.FunctionTrigger(fn, args=None, kwargs=None, trigger=(1, 'iteration'))#

Bases: Trigger

Methods

__init__(fn[, args, kwargs, trigger])

load_state_dict(to_load)

may_fire(iteration, epoch_len)

Flags if the trigger may fire at the current iteration

state_dict()

Parameters:

  • fn (Callable[[...], bool]) –

  • args (Optional[Sequence[Any]]) –

  • kwargs (Optional[Mapping[str, Any]]) –

  • trigger (TriggerLike) –

__call__(manager)#

Call self as a function.

Parameters:

manager (ExtensionsManagerProtocol) –

Return type:

bool

__init__(fn, args=None, kwargs=None, trigger=(1, 'iteration'))#
Parameters:

  • fn (Callable[[...], bool]) –

  • args (Optional[Sequence[Any]]) –

  • kwargs (Optional[Mapping[str, Any]]) –

  • trigger (TriggerLike) –

Return type:

None

load_state_dict(to_load)#
Parameters:

to_load (Dict[str, Any]) –

Return type:

None

may_fire(iteration, epoch_len)#

Flags if the trigger may fire at the current iteration

This must not alter the trigger state

Parameters:

  • iteration (int) –

  • epoch_len (int) –

Return type:

bool

state_dict()#
Return type:

Dict[str, Any]