def get_loaders(train_dir,
train_maskdir,
val_dir,
val_maskdir,
batch_size,
train_transform,
val_transform,
num_workers=4,
pin_memory=True):
train_ds = CarvanaDataset(
image_dir=train_dir,
mask_dir=train_maskdir,
transform=train_transform,
)
train_loader = DataLoader(train_ds,
batch_size=batch_size,
num_workers=num_workers,
pin_memory=pin_memory,
shuffle=True)
val_ds = CarvanaDataset(
image_dir=val_dir,
mask_dir=val_maskdir,
transform=val_transform,
)
val_loader = DataLoader(val_ds,
batch_size=batch_size,
num_workers=num_workers,
pin_memory=pin_memory,
shuffle=False)
return train_loader, val_loader
def get_loaders(
train_dir,
train_maskdir,
val_dir,
val_maskdir,
batch_size,
train_transforms,
val_transforms,
num_workers=4,
pin_memory=True
):
train_ds = CarvanaDataset(train_dir, train_maskdir, train_transforms)
train_loader = DataLoader(
train_ds,
batch_size=batch_size,
num_workers=num_workers,
pin_memory=pin_memory,
shuffle=True
)
val_ds = CarvanaDataset(val_dir, val_maskdir, val_transforms)
val_loader = DataLoader(
val_ds,
batch_size=batch_size,
num_workers=num_workers,
pin_memory=pin_memory,
shuffle=True
)
return train_loader, val_loader
import matplotlib.pyplot as plt
from dataset import CarvanaDataset
random.seed(42)
# %%
image_dir = osp.join("data", "val_set", "images")
mask_dir = osp.join("data", "val_set", "masks")
# ================================================================= #
# Load Data your dataset #
# ================================================================= #
# %% 커스텀 데이터 로드
dataset = CarvanaDataset(image_dir=image_dir,
mask_dir=mask_dir,
transform=None)
# %% 이미지 & 마스크 가져오기
idx = random.randrange(0, len(dataset)) # __len__ 메소드 호출
image, mask = dataset[idx] # __getitem__ 메소드 호출
image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
print(image.shape)
print(mask.shape)
# %% overlay segmented image on top of main image in python
""" Method1
(ref) https://stackoverflow.com/questions/57576686/how-to-overlay-segmented-image-on-top-of-main-image-in-python
"""
def val_dataloader(self):
# OPTIONAL
return DataLoader(CarvanaDataset(folds=self.val_folds),
batch_size=config.VAL_BATCH_SIZE)
def train_dataloader(self):
# REQUIRED
return DataLoader(CarvanaDataset(folds=self.train_folds),
shuffle=True,
batch_size=config.TRAIN_BATCH_SIZE)
parser.add_argument('--momentum', type=float, default=0.5, metavar='M',
help='SGD momentum (default: 0.5)')
parser.add_argument('--no-cuda', action='store_true', default=False,
help='disables CUDA training')
parser.add_argument('--seed', type=int, default=1, metavar='S',
help='random seed (default: 1)')
parser.add_argument('--log-interval', type=int, default=10, metavar='N',
help='how many batches to wait before logging training status')
args = parser.parse_args()
args.cuda = not args.no_cuda and torch.cuda.is_available()
torch.manual_seed(args.seed)
if args.cuda:
torch.cuda.manual_seed(args.seed)
test_ds = CarvanaDataset()
test_ds.initialize(args, phase='test')
kwargs = {'num_workers': 1, 'pin_memory': True} if args.cuda else {}
test_loader = DataLoader(test_ds, batch_size=args.batch_size, drop_last=True, **kwargs)
model = CarvanaFvbNet()
model.load_state_dict(torch.load('./checkpoints/latest_{}.pth'.format(args.which_epoch)))
if args.cuda:
model.cuda()
def test():
model.eval()
print(model)
test_loss = 0
correct = 0
for data, target, dsidx in test_loader:
parser.add_argument('--vis-aimg',
type=int,
default=0,
help='index of the image to visualize activations for')
parser.add_argument('--vis-alayer',
type=int,
default=0,
help='index of the layer to visualize (0-23)')
args = parser.parse_args()
args.cuda = not args.no_cuda and torch.cuda.is_available()
torch.manual_seed(args.seed)
if args.cuda:
torch.cuda.manual_seed(args.seed)
val_ds = CarvanaDataset()
val_ds.initialize(args, phase='val')
kwargs = {'num_workers': 1, 'pin_memory': True} if args.cuda else {}
val_loader = DataLoader(val_ds,
batch_size=args.batch_size,
drop_last=True,
**kwargs)
model = CarvanaFvbNet()
print(model)
print('\nloading model params')
model.load_state_dict(
torch.load('./checkpoints/latest_{}.pth'.format(args.which_epoch)))
if args.cuda:
model.cuda()
print('\nload complete!')