avalanche.evaluation.metrics.StreamBWT

class avalanche.evaluation.metrics.StreamBWT[source]

The StreamBWT metric, emitting the average BWT across all experiences encountered during training.

This plugin metric, computed over all observed experiences during training, is the average over the difference between the last accuracy result obtained on an experience and the accuracy result obtained when first training on that experience.

This metric is computed during the eval phase only.

__init__()

Creates an instance of the StreamForgetting metric.

Methods

__init__()

Creates an instance of the StreamForgetting metric.

after_backward(strategy)

after_eval(strategy)

after_eval_dataset_adaptation(strategy)

after_eval_exp(strategy)

after_eval_forward(strategy)

after_eval_iteration(strategy)

after_forward(strategy)

after_train_dataset_adaptation(strategy)

after_training(strategy)

after_training_epoch(strategy)

after_training_exp(strategy)

after_training_iteration(strategy)

after_update(strategy)

before_backward(strategy)

before_eval(strategy)

before_eval_dataset_adaptation(strategy)

before_eval_exp(strategy)

before_eval_forward(strategy)

before_eval_iteration(strategy)

before_forward(strategy)

before_train_dataset_adaptation(strategy)

before_training(strategy)

before_training_epoch(strategy)

before_training_exp(strategy)

before_training_iteration(strategy)

before_update(strategy)

exp_result([k])

Result for experience defined by a key.

exp_update(k, v[, initial])

Update forgetting metric.

metric_result(strategy)

metric_update(strategy)

reset()

Resets the forgetting metrics.

reset_last()

Resets the last metric value.

result([k])

The average forgetting over all experience.

update(k, v[, initial])

Update forgetting metric.

Attributes