SGD

Arguments

This module implements the simplest form of incremental training, i.e., finetuning.

Classes

class models.sgd.Sgd(backbone, loss, args, transform, dataset=None)

Bases: ContinualModel

Finetuning baseline - simple incremental training.

COMPATIBILITY: List[str] = ['class-il', 'domain-il', 'task-il', 'general-continual']

NAME: str = 'sgd'

observe(inputs, labels, not_aug_inputs, epoch=None, **kwargs)

SGD trains on the current task using the data provided, with no countermeasures to avoid forgetting.