pytorch_pfn_extras.training.extensions.accumulate.max_accumulate.AccumulateBase#

class pytorch_pfn_extras.training.extensions.accumulate.max_accumulate.AccumulateBase(conversion_key_pair, trigger=(1, 'epoch'), distributed=False)#

Bases: ABC, Extension

Methods

__init__(conversion_key_pair[, trigger, ...])

finalize(manager)

Finalizes the extension.

initialize(manager)

Initializes up the manager state.

load_state_dict(to_load)

on_error(manager, exc, tb)

Handles the error raised during training before finalization.

state_dict()

Serializes the extension state.

Attributes

default_name

Default name of the extension.

is_async

name

needs_model_state

priority

trigger

Parameters:

  • conversion_key_pair (Tuple[str, str]) –

  • trigger (Optional[Union[Trigger, Callable[[ExtensionsManagerProtocol], bool], Tuple[float, str]]]) –

  • distributed (bool) –

__call__(manager)#

Call self as a function.

Parameters:

manager (ExtensionsManagerProtocol) –

Return type:

None

__init__(conversion_key_pair, trigger=(1, 'epoch'), distributed=False)#
Parameters:

  • conversion_key_pair (Tuple[str, str]) –

  • trigger (Optional[Union[Trigger, Callable[[ExtensionsManagerProtocol], bool], Tuple[float, str]]]) –

  • distributed (bool) –

Return type:

None

load_state_dict(to_load)#
Parameters:

to_load (Dict[str, Any]) –

Return type:

None

priority: int = 200#
state_dict()#

Serializes the extension state.

It is called when a manager that owns this extension is serialized. It serializes nothing by default.

Return type:

Dict[str, Any]