pytorch_pfn_extras.training.extension.ExtensionEntry#

class pytorch_pfn_extras.training.extension.ExtensionEntry(extension, *, name=None, priority=None, trigger=None, call_before_training=False)#

Bases: object

Extension and options. When name, priority, or trigger is not specified, it is copied from the attributes of the given extension.

Parameters:

  • extension (ExtensionLike) – An extension.

  • name (Optional[str]) – Name of extension.

  • priority (Optional[int]) – Invocation priority of the extension.

  • trigger (Optional[TriggerLike]) – Trigger object that determines when to invoke the extension.

  • call_before_training (bool) – Flag to call extension before training.

See also

pytorch_pfn_extras.training.ExtensionsManager.extend()

Methods

__init__(extension, *[, name, priority, ...])

load_state_dict(to_load)

state_dict()

__init__(extension, *, name=None, priority=None, trigger=None, call_before_training=False)#
Parameters:

  • extension (ExtensionLike) –

  • name (Optional[str]) –

  • priority (Optional[int]) –

  • trigger (Optional[TriggerLike]) –

  • call_before_training (bool) –

Return type:

None

load_state_dict(to_load)#
Parameters:

to_load (Dict[str, Any]) –

Return type:

None

state_dict()#
Return type:

Dict[str, Any]