reset optimizer or not?

[artcmd]

Thanks for sharing this project!

During supconH training, optimizer will be reset every 60 epochs.
https://github.com/tttianhao/CLEAN/blob/main/app/train-supconH.py#L103-L104

# training
for epoch in range(1, epochs + 1):
    if epoch % args.adaptive_rate == 0 and epoch != epochs + 1:
        optimizer = torch.optim.Adam(model.parameters(), lr=lr, betas=(0.9, 0.999))

Could you explain in more detail why this step is needed? Is it possible to skip it?

Thank you in advance!