HAL

Arguments

Options

--hal_lambdafloat

Help: None

• Default: 0.1

--betafloat

Help: None

• Default: 0.5

--gammafloat

Help: None

• Default: 0.1

Rehearsal arguments

Arguments shared by all rehearsal-based methods.

--buffer_sizeint

Help: The size of the memory buffer.

• Default: None

--minibatch_sizeint

Help: The batch size of the memory buffer.

• Default: None

Classes

class models.hal.HAL(backbone, loss, args, transform, dataset=None)

Bases: ContinualModel

Hindsight Anchor Learning.

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

NAME: str = 'hal'

end_task(dataset)

get_anchors(dataset)

static get_parser(parser)

Return type:

ArgumentParser

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