batch_size=100,
shuffle=False,
num_workers=2)
classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse',
'ship', 'truck')
# Model
print('==> Building model..')
checkpoint_dir = 'checkpoints'
if not os.path.exists('./checkpoints'):
os.makedirs('./checkpoints')
if (args.model == 'resnet50'):
net = resnet.ResNet50()
elif (args.model == 'resnet18'):
net = resnet.ResNet18()
net = net.to(device)
if device == 'cuda':
net = torch.nn.DataParallel(net)
cudnn.benchmark = True
print('==> Resuming from checkpoint..')
#assert os.path.isdir(checkpoint_dir), 'Error: no checkpoint directory found!'
resume_file = '{}/{}'.format(checkpoint_dir, args.checkpoint)
assert os.path.isfile(resume_file)
checkpoint = torch.load(resume_file)
net.load_state_dict(checkpoint['net'])
net.eval()
return {'val_loss': avg_loss}
def configure_optimizers(self):
return torch.optim.Adam(self.parameters(), lr=.001)
ap = argparse.ArgumentParser(description="classifier")
ap.add_argument("--model", default="Classifier")
ap.add_argument("--max_epochs", default=10, type=int)
ap.add_argument("--tensorboard_log")
ns = ap.parse_args()
if ns.model == "lenet":
model = lenet.LeNet()
if ns.model == "resnet":
model = resnet.ResNet18()
elif ns.model == "resnetv2":
model = resnetv2.PreActResNet18()
elif ns.model == "densenet":
model = densenet.DenseNet121()
elif ns.model == "resnext":
model = resnext.ResNeXt29_2x64d()
lit_model = Classifier(model)
dataset = datasets.CIFAR10(root='/home/julian/ImageDataSets/CIFAR10',
train=True,
download=False,
transform=transforms.Compose(
[transforms.ToTensor()]))
dataset_size = len(dataset)
print("Dataset size=", dataset_size)
training_size = np.round(dataset_size * 0.9).astype(int)