Engine¶

Events¶

class argus.engine.Events(value)[source]¶

Events that are fired by the argus.engine.Engine during running.

Built-in events:

START: triggered when the engine’s run is started.
COMPLETE: triggered when the engine’s run is completed.
EPOCH_START: triggered when the epoch is started.
EPOCH_COMPLETE: triggered when the epoch is ended.
ITERATION_START: triggered when an iteration is started.
ITERATION_COMPLETE: triggered when the iteration is ended.
CATCH_EXCEPTION: triggered on catching of an exception.

class argus.engine.EventEnum(value)[source]¶: Base class for engine events. User defined custom events should also inherit this class. Example of creating custom events you can find here.

State¶

class argus.engine.State(step_method: Callable[[Any, argus.engine.engine.State], Any], engine: Optional[argus.engine.engine.Engine] = None, phase_states: Optional[Dict[str, argus.engine.engine.State]] = None, **kwargs)[source]¶

A state used to store internal and user-defined variables during a run of argus.engine.Engine. The class is highly inspired by the State from pytorch-ignite.

Parameters

step_method (Callable) – Method of argus.model.Model that takes batch, state and returns step output.
engine (Engine, optional) – argus.engine.Engine that uses this object as a state.
phase_states (dict, optional) – Dictionary with states for each training phase.
**kwargs – Initial attributes of the state.

By default, the state contains the following attributes.

iteration¶

Iteration, the first iteration is 0.

Type: int

epoch¶

Epoch, the first iteration is 0.

Type: int

model¶

argus.Model that uses argus.engine.State.engine and this object as a state.

Type: argus.Model

data_loader¶

A data passed to the argus.engine.Engine.

Type: Iterable, optional

logger¶

Logger.

Type: logging.Logger

exception¶

Catched exception.

Type: BaseException, optional

engine¶

argus.engine.Engine that uses this object as a state.

Type: Engine, optional

phase¶

A phase of training this state was created for. The value takes from the name of the method step_method. If the step_method name ends with _step, the postfix will be removed. For default steps of argus model values are ‘train’ and ‘val’.

Type: str

phase_states¶

Dictionary with states for each training phase.

Type: dict, optional

batch¶

Batch sample from a data loader on the current iteration.

Type: Any

step_output¶

Current output from step_method on current iteration.

Type: Any

metrics¶

Dictionary with metrics values.

Type: dict

stopped¶

Boolean indicates argus.engine.Engine is stopped or not.

Type: bool

update(**kwargs)[source]¶

Update state attributes.

Parameters: **kwargs – Update attributes using kwargs

Engine¶

class argus.engine.Engine(step_method: Callable[[Any, argus.engine.engine.State], Any], phase_states: Optional[Dict[str, argus.engine.engine.State]] = None, **kwargs)[source]¶

Runs step_method over each batch of a data loader with triggering event handlers. The class is highly inspired by the Engine from pytorch-ignite.

Parameters

step_method (Callable) – Method of argus.model.Model that takes batch, state and returns step output.
phase_states (dict, optional) – Dictionary with states for each training phase.
**kwargs – Initial attributes of the state.

state¶

Stores internal and user-defined variables during a run of the engine.

Type: State

step_method¶

Method of argus.model.Model that takes batch, state and returns step output.

Type: Callable

event_handlers (dict of EventEnum: list): Dictionary that stores event handlers.

add_event_handler(event: argus.engine.engine.EventEnum, handler: Callable, *args, **kwargs)[source]¶

Add an event handler to be executed when the event is triggered.

Parameters

event (EventEnum) – An event that will be associated with the handler.
handler (Callable) – A callable handler that will be executed on the event. The handler should take argus.engine.State as the first argument.
*args – optional args arguments to be passed to the handler.
**kwargs – optional kwargs arguments to be passed to the handler.

raise_event(event: argus.engine.engine.EventEnum)[source]¶

Execute all the handlers associated with the given event.

Parameters: event (EventEnum) – An event that will be triggered.

run(data_loader: Iterable, start_epoch: int = 0, end_epoch: int = 1) → argus.engine.engine.State[source]¶

Run step_method on each batch from data loader end_epoch - start_epoch times.

Parameters

data_loader (Iterable) – An iterable collection that returns batches.
start_epoch (int) – The first epoch number.
end_epoch (int) – One above the largest epoch number.

Returns

An engine state.

Return type

State

Metrics

Utils