GEM baseline in class-IL for GCGL has wrong indentation for optimizer step?

[WeiWeic6222848]

GEM baseline in class-IL for GCGL has wrong indentation for optimizer step?

[QueuQ]

as seen in the next code fragment, line 212, optimizer.step() is outside of the for loop, which iterates over the current task.

CGLB/GCGL/Baselines/gem_model.py

Lines 178 to 212 in 3e0debf

	for batch_id, batch_data in enumerate(data_loader[task_i]):
	smiles, bg, labels, masks = batch_data
	bg = bg.to(f"cuda:{args['gpu']}")
	labels, masks = labels.cuda(), masks.cuda()
	logits = predict(args, self.net, bg)
	# class balance
	n_per_cls = [(labels == j).sum() for j in clss]
	loss_w_ = [1. / max(i, 1) for i in n_per_cls]
	loss_w_ = torch.tensor(loss_w_).to(device='cuda:{}'.format(args['gpu']))
	# labels= labels.long()
	for i, c in enumerate(clss):
	labels[labels == c] = i
	# Mask non-existing labels
	loss = loss_criterion(logits[:, clss], labels.long(), weight=loss_w_).float()
	self.optimizer.zero_grad()
	loss.backward()
	# check if gradient violates constraints
	if len(self.observed_tasks) > 1:
	# copy gradient
	store_grad(self.net.parameters, self.grads, self.grad_dims, task_i)
	indx = torch.cuda.LongTensor(self.observed_tasks[:-1])
	dotp = torch.mm(self.grads[:, task_i].unsqueeze(0),
	self.grads.index_select(1, indx))
	if (dotp < 0).sum() != 0:
	project2cone2(self.grads[:, task_i].unsqueeze(1),
	self.grads.index_select(1, indx), self.margin)
	# copy gradients back
	overwrite_grad(self.net.parameters, self.grads[:, task_i],
	self.grad_dims)
	self.optimizer.step()

while in the same file, in line 296 and line 127, this step is inside the for loop.

Thanks for pointing out this issue. It should be a mistake, and we have corrected it and updated it. We tested the result of GEM with the updated code (on Aromaticity-CL, the only class-IL dataset), and got AP 5.2±1.1 AF -57.6±3.4. The AP decreases a little, and the AF increases. The previous code would skip some batches of data in each epoch, but since shuffle is used for the data loader, all data should get chance to be trained after multiple epochs. Now the corrected code would ensure that all batches are used in each epoch.

[WeiWeic6222848]

Thank you for the update