Skip to content

Commit

Permalink
correct ordlayer
Browse files Browse the repository at this point in the history
  • Loading branch information
@dontLoveBugs
dontLoveBugs committed Mar 3, 2019
1 parent 0e982db commit ff33276
Show file tree
Hide file tree
Showing 6 changed files with 154 additions and 209 deletions.
4 changes: 2 additions & 2 deletions dataloaders/path.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,9 +10,9 @@ class Path(object):
@staticmethod
def db_root_dir(database):
if database == 'nyu':
return '/home/data/UnsupervisedDepth/nyudepthv2'
return '/home/data/model/wangxin/nyudepthv2'
elif database == 'kitti':
return '/home/data/UnsupervisedDepth/KITTI'
return '/home/data/UnsupervisedDepth/wangixn/KITTI'
else:
print('Database {} not available.'.format(database))
raise NotImplementedError
1 change: 1 addition & 0 deletions main.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -272,6 +272,7 @@ def validate(val_loader, model, epoch, logger):
end = time.time()

skip = len(val_loader) // 9 # save images every skip iters
img_list = []

for i, (input, target) in enumerate(val_loader):

Expand Down
104 changes: 5 additions & 99 deletions network/DORN_kitti.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,8 @@
import collections
import math

from network.backbone import resnet101


def weights_init(modules, type='xavier'):
m = modules
Expand Down Expand Up @@ -181,22 +183,6 @@ def forward(self, x):
N, C, H, W = x.size()
ord_num = C // 2

# if torch.cuda.is_available():
# decode_label = torch.zeros((N, 1, H, W), dtype=torch.float32).cuda()
# ord_labels = torch.zeros((N, C // 2, H, W), dtype=torch.float32).cuda()
# else:
# decode_label = torch.zeros((N, 1, H, W), dtype=torch.float32)
# ord_labels = torch.zeros((N, C // 2, H, W), dtype=torch.float32)
# print('#1 decode size:', decode_label.size())
# ord_num = C // 2
# for i in range(ord_num):
# ord_i = x[:, 2 * i:2 * i + 2, :, :]
# ord_i = nn.functional.softmax(ord_i, dim=1) # compute P(w, h) in paper
# ord_i = ord_i[:, 1, :, :]
# ord_labels[:, i, :, :] = ord_i
# # print('ord_i >= 0.5 size:', (ord_i >= 0.5).size())
# decode_label += (ord_i >= 0.5).view(N, 1, H, W).float() # sum(n(p_k >= 0.5))

"""
replace iter with matrix operation
fast speed methods
Expand All @@ -214,97 +200,17 @@ def forward(self, x):

ord_c1 = ord_c[:, 1, :].clone()
ord_c1 = ord_c1.view(-1, ord_num, H, W)
decode_c = torch.sum(ord_c1, dim=1).view(-1, 1, H, W)
decode_c = torch.sum((ord_c1 > 0.5), dim=1).view(-1, 1, H, W)
return decode_c, ord_c1


class ResNet(nn.Module):
def __init__(self, in_channels=3, pretrained=True, freeze=True):
super(ResNet, self).__init__()
pretrained_model = torchvision.models.__dict__['resnet{}'.format(101)](pretrained=pretrained)

self.channel = in_channels

self.conv1 = nn.Sequential(collections.OrderedDict([
('conv1_1', nn.Conv2d(self.channel, 64, kernel_size=3, stride=2, padding=1, bias=False)),
('bn1_1', nn.BatchNorm2d(64)),
('relu1_1', nn.ReLU(inplace=True)),
('conv1_2', nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1, bias=False)),
('bn_2', nn.BatchNorm2d(64)),
('relu1_2', nn.ReLU(inplace=True)),
('conv1_3', nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1, bias=False)),
('bn1_3', nn.BatchNorm2d(128)),
('relu1_3', nn.ReLU(inplace=True))
]))

# print(pretrained_model._modules['layer1'][0].conv1)

self.maxpool = pretrained_model._modules['maxpool']
self.layer1 = pretrained_model._modules['layer1']
self.layer1[0].conv1 = nn.Conv2d(128, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
self.layer1[0].downsample[0] = nn.Conv2d(128, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)

self.layer2 = pretrained_model._modules['layer2']

self.layer3 = pretrained_model._modules['layer3']
self.layer3[0].conv2 = nn.Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
self.layer3[0].downsample[0] = nn.Conv2d(512, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)

self.layer4 = pretrained_model._modules['layer4']
self.layer4[0].conv2 = nn.Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
self.layer4[0].downsample[0] = nn.Conv2d(1024, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)

# clear memory
del pretrained_model

if pretrained:
weights_init(self.conv1, type='kaiming')
weights_init(self.layer1[0].conv1, type='kaiming')
weights_init(self.layer1[0].downsample[0], type='kaiming')
weights_init(self.layer3[0].conv2, type='kaiming')
weights_init(self.layer3[0].downsample[0], type='kaiming')
weights_init(self.layer4[0].conv2, 'kaiming')
weights_init(self.layer4[0].downsample[0], 'kaiming')
else:
weights_init(self.modules(), type='kaiming')

if freeze:
self.freeze()

def forward(self, x):
# print(pretrained_model._modules)

x = self.conv1(x)

# print('conv1:', x.size())

x = self.maxpool(x)

# print('pool:', x.size())

x1 = self.layer1(x)
# print('layer1 size:', x1.size())
x2 = self.layer2(x1)
# print('layer2 size:', x2.size())
x3 = self.layer3(x2)
# print('layer3 size:', x3.size())
x4 = self.layer4(x3)
# print('layer4 size:', x4.size())
return x4

def freeze(self):
for m in self.modules():
if isinstance(m, nn.BatchNorm2d):
m.eval()


class DORN(nn.Module):
def __init__(self, output_size=(257, 353), channel=3, pretrained=True, freeze=True):
super(DORN, self).__init__()

self.output_size = output_size
self.channel = channel
self.feature_extractor = ResNet(in_channels=channel, pretrained=pretrained, freeze=freeze)
self.feature_extractor = resnet101(pretrained=pretrained)
self.aspp_module = SceneUnderstandingModule()
self.orl = OrdinalRegressionLayer()

Expand Down Expand Up @@ -334,7 +240,7 @@ def get_10x_lr_params(self):
# os.environ["CUDA_VISIBLE_DEVICES"] = "1" # 默认使用GPU 0

if __name__ == "__main__":
model = DORN()
model = DORN(pretrained=False)
model = model.cuda()
model.eval()
image = torch.randn(1, 3, 385, 513)
Expand Down
111 changes: 7 additions & 104 deletions network/DORN_nyu.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,8 @@
import collections
import math

from network.backbone import resnet101


def weights_init(modules, type='xavier'):
m = modules
Expand Down Expand Up @@ -180,22 +182,6 @@ def forward(self, x):
N, C, H, W = x.size()
ord_num = C // 2

# if torch.cuda.is_available():
# decode_label = torch.zeros((N, 1, H, W), dtype=torch.float32).cuda()
# ord_labels = torch.zeros((N, C // 2, H, W), dtype=torch.float32).cuda()
# else:
# decode_label = torch.zeros((N, 1, H, W), dtype=torch.float32)
# ord_labels = torch.zeros((N, C // 2, H, W), dtype=torch.float32)
# print('#1 decode size:', decode_label.size())
# ord_num = C // 2
# for i in range(ord_num):
# ord_i = x[:, 2 * i:2 * i + 2, :, :]
# ord_i = nn.functional.softmax(ord_i, dim=1) # compute P(w, h) in paper
# ord_i = ord_i[:, 1, :, :]
# ord_labels[:, i, :, :] = ord_i
# # print('ord_i >= 0.5 size:', (ord_i >= 0.5).size())
# decode_label += (ord_i >= 0.5).view(N, 1, H, W).float() # sum(n(p_k >= 0.5))

"""
replace iter with matrix operation
fast speed methods
Expand All @@ -213,102 +199,19 @@ def forward(self, x):

ord_c1 = ord_c[:, 1, :].clone()
ord_c1 = ord_c1.view(-1, ord_num, H, W)
decode_c = torch.sum(ord_c1, dim=1).view(-1, 1, H, W)
print('ord > 0.5 size:', (ord_c1 > 0.5).size())
decode_c = torch.sum((ord_c1 > 0.5), dim=1).view(-1, 1, H, W)
# decode_c = torch.sum(ord_c1, dim=1).view(-1, 1, H, W)
return decode_c, ord_c1


class ResNet(nn.Module):
def __init__(self, in_channels=3, pretrained=True, freeze=True):
super(ResNet, self).__init__()
pretrained_model = torchvision.models.__dict__['resnet{}'.format(101)](pretrained=pretrained)

self.channel = in_channels

self.conv1 = nn.Sequential(collections.OrderedDict([
('conv1_1', nn.Conv2d(self.channel, 64, kernel_size=3, stride=2, padding=1, bias=False)),
('bn1_1', nn.BatchNorm2d(64)),
('relu1_1', nn.ReLU(inplace=True)),
('conv1_2', nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1, bias=False)),
('bn_2', nn.BatchNorm2d(64)),
('relu1_2', nn.ReLU(inplace=True)),
('conv1_3', nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1, bias=False)),
('bn1_3', nn.BatchNorm2d(128)),
('relu1_3', nn.ReLU(inplace=True))
]))

self.bn1 = nn.BatchNorm2d(128)

# print(pretrained_model._modules['layer1'][0].conv1)

self.relu = pretrained_model._modules['relu']
self.maxpool = pretrained_model._modules['maxpool']
self.layer1 = pretrained_model._modules['layer1']
self.layer1[0].conv1 = nn.Conv2d(128, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
self.layer1[0].downsample[0] = nn.Conv2d(128, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)

self.layer2 = pretrained_model._modules['layer2']

self.layer3 = pretrained_model._modules['layer3']
self.layer3[0].conv2 = nn.Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
self.layer3[0].downsample[0] = nn.Conv2d(512, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)

self.layer4 = pretrained_model._modules['layer4']
self.layer4[0].conv2 = nn.Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
self.layer4[0].downsample[0] = nn.Conv2d(1024, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)

# clear memory
del pretrained_model

if pretrained:
weights_init(self.conv1, type='kaiming')
weights_init(self.layer1[0].conv1, type='kaiming')
weights_init(self.layer1[0].downsample[0], type='kaiming')
weights_init(self.layer3[0].conv2, type='kaiming')
weights_init(self.layer3[0].downsample[0], type='kaiming')
weights_init(self.layer4[0].conv2, 'kaiming')
weights_init(self.layer4[0].downsample[0], 'kaiming')
else:
weights_init(self.modules(), type='kaiming')

if freeze:
self.freeze()

def forward(self, x):
# print(pretrained_model._modules)

x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)

# print('conv1:', x.size())

x = self.maxpool(x)

# print('pool:', x.size())

x1 = self.layer1(x)
# print('layer1 size:', x1.size())
x2 = self.layer2(x1)
# print('layer2 size:', x2.size())
x3 = self.layer3(x2)
# print('layer3 size:', x3.size())
x4 = self.layer4(x3)
# print('layer4 size:', x4.size())
return x4

def freeze(self):
for m in self.modules():
if isinstance(m, nn.BatchNorm2d):
m.eval()


class DORN(nn.Module):
def __init__(self, output_size=(257, 353), channel=3, pretrained=True, freeze=True):
super(DORN, self).__init__()

self.output_size = output_size
self.channel = channel
self.feature_extractor = ResNet(in_channels=channel, pretrained=pretrained, freeze=freeze)
self.feature_extractor = resnet101(pretrained=pretrained)
self.aspp_module = SceneUnderstandingModule()
self.orl = OrdinalRegressionLayer()

Expand Down Expand Up @@ -338,7 +241,7 @@ def get_10x_lr_params(self):
# os.environ["CUDA_VISIBLE_DEVICES"] = "1" # 默认使用GPU 0

if __name__ == "__main__":
model = DORN()
model = DORN(pretrained=False)
model = model.cuda()
model.eval()
image = torch.randn(1, 3, 257, 353)
Expand Down
Loading

0 comments on commit ff33276

Please sign in to comment.