def __init__(self, in_planes, planes, stride=1):
super(Bottleneck, self).__init__()
self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=1, bias=False)
self.bn1 = nn.BatchNorm2d(planes)
#self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride, padding=1, bias=False)
self.conv2 = deform_conv.DeformConv2D(planes,
planes,
kernel_size=3,
padding=1)
self.bn2 = nn.BatchNorm2d(planes)
#self.conv3 = deform_conv.DeformConv2D(planes, self.expansion*planes, kernel_size=1, bias=False)
self.conv3 = nn.Conv2d(planes,
self.expansion * planes,
kernel_size=1,
bias=False)
self.bn3 = nn.BatchNorm2d(self.expansion * planes)
self.shortcut = nn.Sequential()
#print("shortcut")
if stride != 1 or in_planes != self.expansion * planes:
#print("stride: ", stride, " in_planes: ", in_planes)
self.shortcut = nn.Sequential(
nn.Conv2d(in_planes,
self.expansion * planes,
kernel_size=1,
stride=stride,
bias=False), nn.BatchNorm2d(self.expansion * planes))
def __init__(self):
super(MyAttenNet, self).__init__()
self.base_model = nn.Sequential(
nn.Conv2d(1, 64, kernel_size=4, padding=2),
nn.BatchNorm2d(64),
nn.PReLU(),
nn.MaxPool2d(kernel_size=2),
nn.Dropout2d(0.1),
deform_conv.DeformConv2D(64, 128, kernel_size=4, padding=1),
nn.BatchNorm2d(128),
nn.PReLU(),
nn.MaxPool2d(kernel_size=2),
nn.Dropout2d(0.3),
#deform_conv.DeformConv2D(128, 128, kernel_size=3, padding=1),
)
self.atten = nn.Sequential(nn.Conv2d(128, 1, 1, 1),
nn.Softplus(beta=1, threshold=20))
self.fc = nn.Sequential(
nn.Linear(7 * 7 * 128, 512),
nn.PReLU(),
nn.Dropout(0.5),
nn.Linear(512, 64),
nn.PReLU(),
nn.Dropout(0.5),
nn.Linear(64, 10),
)
def __init__(self, in_planes, planes, stride=1):
super(BasicBlock, self).__init__()
self.conv1 = nn.Conv2d(in_planes,
planes,
kernel_size=3,
stride=stride,
padding=1,
bias=False)
self.bn1 = nn.BatchNorm2d(planes)
#self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=1, padding=1, bias=False)
self.conv2 = deform_conv.DeformConv2D(planes,
planes,
stride=stride,
kernel_size=3,
bias=False)
self.bn2 = nn.BatchNorm2d(planes)
self.shortcut = nn.Sequential()
if stride != 1 or in_planes != self.expansion * planes:
self.shortcut = nn.Sequential(
nn.Conv2d(in_planes,
self.expansion * planes,
kernel_size=1,
stride=stride,
bias=False), nn.BatchNorm2d(self.expansion * planes))
def __init__(self):
super(DeformNet, self).__init__()
self.conv1 = nn.Conv2d(3, 32, kernel_size=3, padding=1, bias = False)
self.bn1 = nn.BatchNorm2d(32)
self.conv2 = nn.Conv2d(32, 64, kernel_size=3, padding=1, bias = False)
self.bn2 = nn.BatchNorm2d(64)
self.conv3 = nn.Conv2d(64, 128, kernel_size=3, padding=1, bias = False)
self.bn3 = nn.BatchNorm2d(128)
self.conv4 = deform_conv.DeformConv2D(128, 128, kernel_size=3, padding=1, bias = False)
self.bn4 = nn.BatchNorm2d(128)
#self.classifier = nn.Linear(128, 10)
self.classifier = nn.Linear(128, 10)
def __init__(self, in_planes, planes, stride=1):
#print("Bottleneck2")
super(Bottleneck2, self).__init__()
#cannot apply deformable conv to conv1 and conv3 becasue the kernel size is 1
self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=1, bias=False)
self.bn1 = nn.BatchNorm2d(planes)
self.offsets = nn.Conv2d(planes,
18,
kernel_size=3,
padding=1,
stride=stride)
self.conv2 = deform_conv.DeformConv2D(planes,
planes,
kernel_size=3,
padding=1)
#self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride, padding=1, bias=False)
self.bn2 = nn.BatchNorm2d(planes)
self.conv3 = nn.Conv2d(planes,
self.expansion * planes,
kernel_size=1,
bias=False)
self.bn3 = nn.BatchNorm2d(self.expansion * planes)
self.shortcut = nn.Sequential()
#print("shortcut")
if stride != 1 or in_planes != self.expansion * planes:
#print("stride: ", stride, " in_planes: ", in_planes)
self.shortcut = nn.Sequential(
nn.Conv2d(in_planes,
self.expansion * planes,
kernel_size=1,
stride=stride,
bias=False), nn.BatchNorm2d(self.expansion * planes))