print(("=> loaded checkpoint '{}' (epoch {})".format(
args.evaluate, checkpoint['epoch'])))
else:
print(("=> no checkpoint found at '{}'".format(args.resume)))
# dataset
train_transform = T.Compose([
train_augmentation,
ut_transforms.Stack(roll=args.arch == 'BNInception'),
ut_transforms.ToTorchFormatTensor(div=args.arch != 'BNInception'),
ut_transforms.IdentityTransform(),
])
test_transform = T.Compose([
ut_transforms.GroupScale(int(scale_size)),
ut_transforms.GroupCenterCrop(crop_size),
ut_transforms.Stack(roll=args.arch == 'BNInception'),
ut_transforms.ToTorchFormatTensor(div=args.arch != 'BNInception'),
ut_transforms.IdentityTransform(),
])
train_dataset = TSNDataSet(num_segments=args.num_segments,
dataset=args.dataset,
new_length=data_length,
modality=args.modality,
transform=train_transform,
split=args.split,
train=True)
test_dataset = TSNDataSet(num_segments=args.num_segments,
dataset=args.dataset,
def main():
global args, best_mIoU
args = parser.parse_args()
os.environ["CUDA_VISIBLE_DEVICES"] = ','.join(str(gpu) for gpu in args.gpus)
args.gpus = len(args.gpus)
if args.no_partialbn:
sync_bn.Synchronize.init(args.gpus)
if args.dataset == 'VOCAug' or args.dataset == 'VOC2012' or args.dataset == 'COCO':
num_class = 21
ignore_label = 255
scale_series = [10, 20, 30, 60]
elif args.dataset == 'cityscapes':
num_class = 19
ignore_label = 0
scale_series = [15, 30, 45, 90]
else:
raise ValueError('Unknown dataset ' + args.dataset)
model = models.FCN(num_class, base_model=args.arch, dropout=args.dropout, partial_bn=not args.no_partialbn)
input_mean = model.input_mean
input_std = model.input_std
policies = model.get_optim_policies()
model = torch.nn.DataParallel(model, device_ids=range(args.gpus)).cuda()
if args.resume:
if os.path.isfile(args.resume):
print(("=> loading checkpoint '{}'".format(args.resume)))
checkpoint = torch.load(args.resume)
args.start_epoch = checkpoint['epoch']
best_mIoU = checkpoint['best_mIoU']
torch.nn.Module.load_state_dict(model, checkpoint['state_dict'])
print(("=> loaded checkpoint '{}' (epoch {})".format(args.evaluate, checkpoint['epoch'])))
else:
print(("=> no checkpoint found at '{}'".format(args.resume)))
if args.weight:
if os.path.isfile(args.weight):
print(("=> loading initial weight '{}'".format(args.weight)))
checkpoint = torch.load(args.weight)
torch.nn.Module.load_state_dict(model, checkpoint['state_dict'])
else:
print(("=> no model file found at '{}'".format(args.weight)))
cudnn.benchmark = True
cudnn.fastest = True
# Data loading code
train_loader = torch.utils.data.DataLoader(
getattr(ds, args.dataset + 'DataSet')(data_list=args.train_list, transform=torchvision.transforms.Compose([
tf.GroupRandomHorizontalFlip(),
tf.GroupRandomScale(size=(0.5, 2.0), interpolation=(cv2.INTER_LINEAR, cv2.INTER_NEAREST)),
tf.GroupRandomCrop(size=args.train_size),
tf.GroupRandomPad(size=args.train_size, padding=(input_mean, (ignore_label, ))),
tf.GroupRandomRotation(degree=(-10, 10), interpolation=(cv2.INTER_LINEAR, cv2.INTER_NEAREST), padding=(input_mean, (ignore_label, ))),
tf.GroupRandomBlur(applied=(True, False)),
tf.GroupNormalize(mean=(input_mean, (0, )), std=(input_std, (1, ))),
])), batch_size=args.batch_size, shuffle=True, num_workers=args.workers, pin_memory=True, drop_last=True)
val_loader = torch.utils.data.DataLoader(
getattr(ds, args.dataset + 'DataSet')(data_list=args.val_list, transform=torchvision.transforms.Compose([
tf.GroupCenterCrop(size=args.test_size),
tf.GroupConcerPad(size=args.test_size, padding=(input_mean, (ignore_label, ))),
tf.GroupNormalize(mean=(input_mean, (0, )), std=(input_std, (1, ))),
])), batch_size=args.batch_size * 3, shuffle=False, num_workers=args.workers, pin_memory=True)
# define loss function (criterion) optimizer and evaluator
criterion = torch.nn.NLLLoss(ignore_index=ignore_label).cuda()
for group in policies:
print(('group: {} has {} params, lr_mult: {}, decay_mult: {}'.format(group['name'], len(group['params']), group['lr_mult'], group['decay_mult'])))
optimizer = torch.optim.SGD(policies, args.lr, momentum=args.momentum, weight_decay=args.weight_decay)
evaluator = EvalSegmentation(num_class, ignore_label)
if args.evaluate:
validate(val_loader, model, criterion, 0, evaluator)
return
for epoch in range(args.start_epoch, args.epochs):
adjust_learning_rate(optimizer, epoch, args.lr_steps)
# train for one epoch
train(train_loader, model, criterion, optimizer, epoch)
# evaluate on validation set
if (epoch + 1) % args.eval_freq == 0 or epoch == args.epochs - 1:
mIoU = validate(val_loader, model, criterion, (epoch + 1) * len(train_loader), evaluator)
# remember best mIoU and save checkpoint
is_best = mIoU > best_mIoU
best_mIoU = max(mIoU, best_mIoU)
save_checkpoint({
'epoch': epoch + 1,
'arch': args.arch,
'state_dict': model.state_dict(),
'best_mIoU': best_mIoU,
}, is_best)
def __call__(self, data):
return data*2 - 1.0
train_transforms = T.Compose([
ut_transforms.GroupScale(256), # resize smaller edge to 256
ut_transforms.GroupRandomCrop(224), # randomlly crop a 224x224 patch
ut_transforms.GroupRandomHorizontalFlip(),
ut_transforms.GroupToTensor(),
ut_transforms.StackTensor(),
I3Dscale()
])
val_transforms = T.Compose([
ut_transforms.GroupScale(256),
ut_transforms.GroupCenterCrop(256), # full convolution in test time
ut_transforms.GroupToTensor(),
ut_transforms.StackTensor(),
I3Dscale()
])
train_dataset = Consecutive(dataset=args.dataset, train=True, transform=train_transforms) #default 64 frames
val_dataset = Consecutive(dataset=args.dataset, train=False, transform=val_transforms, test_mode='else') # just load a 64 frames to test
train_loader = DataLoader(
train_dataset, batch_size=args.batch_size,
shuffle=True, num_workers=args.workers,
pin_memory=True)
# 暂时只支持RGB frames 和kinetics\ucf101数据集
train_transforms = T.Compose([
ut_transforms.GroupRandomScale(
size_low=256,
size_high=320), # randomly resize smaller edge to [256, 320]
ut_transforms.GroupRandomCrop(224), # randomlly crop a 224x224 patch
ut_transforms.GroupToTensor(),
# ut_transforms.GroupNormalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
ut_transforms.StackTensor()
])
val_transforms = T.Compose([
ut_transforms.GroupScale(256), # scale to 256 and do fully-convolutional
ut_transforms.GroupCenterCrop(256),
ut_transforms.GroupToTensor(),
# ut_transforms.GroupNormalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
ut_transforms.StackTensor()
])
train_dataset = Consecutive(
dataset=args.dataset, train=True, interval=2,
transform=train_transforms) #default 64/2 = 32 frames
val_dataset = Consecutive(dataset=args.dataset,
train=False,
interval=2,
transform=val_transforms,
test_mode='else') # also 32 frames
train_loader = DataLoader(train_dataset,
def __call__(self, data):
return data*2 - 1.0
train_transforms = T.Compose([
ut_transforms.GroupScale(256), # resize smaller edge to 256
ut_transforms.GroupRandomCrop(224), # randomlly crop a 224x224 patch
ut_transforms.GroupRandomHorizontalFlip(),
ut_transforms.GroupToTensor(),
ut_transforms.StackTensor(),
I3Dscale()
])
val_transforms = T.Compose([
ut_transforms.GroupScale(256),
ut_transforms.GroupCenterCrop(224), # center crop 224x224 patch
ut_transforms.GroupToTensor(),
ut_transforms.StackTensor(),
I3Dscale()
])
train_dataset = Consecutive(dataset=args.dataset, train=True, transform=train_transforms) #default 64 frames
val_dataset = Consecutive(dataset=args.dataset, train=False, transform=val_transforms, test_mode='else')
train_loader = DataLoader(
train_dataset, batch_size=args.batch_size,
shuffle=True, num_workers=args.workers,
pin_memory=True)
val_loader = DataLoader(
val_dataset, batch_size=1,