def main_with_centerloss(model, epoch):
"""
train model
:param model:
:param epoch:
:param data_name: ISIC/SD198
:return:
"""
criterion_xent = nn.CrossEntropyLoss()
criterion_cent = CenterLoss(num_classes=40, feat_dim=1024)
optimizer_model = optim.SGD(model.parameters(), lr=0.01, momentum=0.9, weight_decay=1e-4)
optimizer_centloss = optim.SGD(criterion_cent.parameters(), lr=0.5)
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_model, step_size=80, gamma=0.1)
trainloader, valloader, testloader = data_loader.load_garbage_classification_data()
dataloaders = {
'train': trainloader,
'val': valloader,
'test': testloader,
}
train_model(model=model, dataloaders=dataloaders, criterion_xent=criterion_xent, criterion_cent=criterion_cent,
optimizer_model=optimizer_model, optimizer_centloss=optimizer_centloss, scheduler=exp_lr_scheduler,
num_epochs=epoch, inference=False)
def train_garbage_classification(model, epoch):
"""
train deep models on garbage classification
:param model:
:param epoch:
:return:
"""
criterion = nn.CrossEntropyLoss()
optimizer_ft = optim.SGD(model.parameters(),
lr=cfg['init_lr'],
momentum=0.9,
weight_decay=cfg['weight_decay'])
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_ft,
step_size=cfg['lr_decay_step'],
gamma=0.1)
print('start loading GarbageDataset...')
trainloader, valloader, testloader = data_loader.load_garbage_classification_data(
)
dataloaders = {
'train': trainloader,
'val': valloader,
'test': testloader,
}
train_model(model=model,
dataloaders=dataloaders,
criterion=criterion,
optimizer=optimizer_ft,
scheduler=exp_lr_scheduler,
num_epochs=epoch,
inference=False)