# Copyright 2020-present, Pietro Buzzega, Matteo Boschini, Angelo Porrello, Davide Abati, Simone Calderara.
# All rights reserved.
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.
import torch
from models.utils.continual_model import ContinualModel
from utils.args import add_rehearsal_args, ArgumentParser
from utils.gss_buffer import Buffer as Buffer
[docs]
class Gss(ContinualModel):
"""Gradient based sample selection for online continual learning."""
NAME = 'gss'
COMPATIBILITY = ['class-il', 'domain-il', 'task-il', 'general-continual']
[docs]
@staticmethod
def get_parser(parser) -> ArgumentParser:
add_rehearsal_args(parser)
parser.add_argument('--batch_num', type=int, default=1,
help='Number of batches extracted from the buffer.')
parser.add_argument('--gss_minibatch_size', type=int, default=None,
help='The batch size of the gradient comparison.')
return parser
def __init__(self, backbone, loss, args, transform, dataset=None):
super(Gss, self).__init__(backbone, loss, args, transform, dataset=dataset)
self.buffer = Buffer(self.args.buffer_size, self.device,
self.args.gss_minibatch_size if
self.args.gss_minibatch_size is not None
else self.args.minibatch_size, self)
self.alj_nepochs = self.args.batch_num
[docs]
def get_grads(self, inputs, labels):
self.net.eval()
self.opt.zero_grad()
outputs = self.net(inputs)
loss = self.loss(outputs, labels)
loss.backward()
grads = self.net.get_grads().clone().detach()
self.opt.zero_grad()
self.net.train()
if len(grads.shape) == 1:
grads = grads.unsqueeze(0)
return grads
[docs]
def observe(self, inputs, labels, not_aug_inputs, epoch=None):
real_batch_size = inputs.shape[0]
self.buffer.drop_cache()
self.buffer.reset_fathom()
for _ in range(self.alj_nepochs):
self.opt.zero_grad()
if not self.buffer.is_empty():
buf_inputs, buf_labels = self.buffer.get_data(
self.args.minibatch_size, transform=self.transform)
tinputs = torch.cat((inputs, buf_inputs))
tlabels = torch.cat((labels, buf_labels))
else:
tinputs = inputs
tlabels = labels
outputs = self.net(tinputs)
loss = self.loss(outputs, tlabels)
loss.backward()
self.opt.step()
self.buffer.add_data(examples=not_aug_inputs,
labels=labels[:real_batch_size])
return loss.item()