def __init__(self, num_classes, loss={'softmax'}, aligned=False, **kwargs):
super(ResNet50, self).__init__()
self.loss = loss
resnet50 = torchvision.models.resnet50(pretrained=True)
self.base = nn.Sequential(*list(resnet50.children())[:-2])
self.classifier = nn.Linear(2048, num_classes)
self.s_classifier = AngleLinear(2048, num_classes)
self.feat_dim = 2048 # feature dimension
self.aligned = aligned
if self.aligned:
self.bn = nn.BatchNorm2d(2048)
self.relu = nn.ReLU(inplace=True)
self.horizon_pool = HorizontalMaxPool2d()
self.conv1 = nn.Conv2d(2048, 128, kernel_size=1, stride=1, padding=0, bias=True)
def __init__(self,
num_classes,
loss={'softmax'},
num_groups=3,
aligned=False,
**kwargs):
super(ShuffleNet, self).__init__()
self.loss = loss
self.conv1 = nn.Sequential(
nn.Conv2d(3, 24, 3, stride=2, padding=1, bias=False),
nn.BatchNorm2d(24),
nn.ReLU(),
nn.MaxPool2d(3, stride=2, padding=1),
)
self.stage2 = nn.Sequential(
Bottleneck(24, cfg[num_groups][0], 2, num_groups),
Bottleneck(cfg[num_groups][0], cfg[num_groups][0], 1, num_groups),
Bottleneck(cfg[num_groups][0], cfg[num_groups][0], 1, num_groups),
Bottleneck(cfg[num_groups][0], cfg[num_groups][0], 1, num_groups),
)
self.stage3 = nn.Sequential(
Bottleneck(cfg[num_groups][0], cfg[num_groups][1], 2, num_groups),
Bottleneck(cfg[num_groups][1], cfg[num_groups][1], 1, num_groups),
Bottleneck(cfg[num_groups][1], cfg[num_groups][1], 1, num_groups),
Bottleneck(cfg[num_groups][1], cfg[num_groups][1], 1, num_groups),
Bottleneck(cfg[num_groups][1], cfg[num_groups][1], 1, num_groups),
Bottleneck(cfg[num_groups][1], cfg[num_groups][1], 1, num_groups),
Bottleneck(cfg[num_groups][1], cfg[num_groups][1], 1, num_groups),
Bottleneck(cfg[num_groups][1], cfg[num_groups][1], 1, num_groups),
)
self.stage4 = nn.Sequential(
Bottleneck(cfg[num_groups][1], cfg[num_groups][2], 2, num_groups),
Bottleneck(cfg[num_groups][2], cfg[num_groups][2], 1, num_groups),
Bottleneck(cfg[num_groups][2], cfg[num_groups][2], 1, num_groups),
Bottleneck(cfg[num_groups][2], cfg[num_groups][2], 1, num_groups),
)
self.classifier = nn.Linear(cfg[num_groups][2], num_classes)
self.feat_dim = cfg[num_groups][2]
self.aligned = aligned
self.horizon_pool = HorizontalMaxPool2d()
def __init__(self, num_classes, loss={'softmax'}, aligned=False, **kwargs):
super(DenseNet121, self).__init__()
self.loss = loss
densenet121 = torchvision.models.densenet121(pretrained=True)
self.base = densenet121.features
self.classifier = nn.Linear(1024, num_classes)
self.feat_dim = 1024 # feature dimension
self.aligned = aligned
self.horizon_pool = HorizontalMaxPool2d()
if self.aligned:
self.bn = nn.BatchNorm2d(1024)
self.relu = nn.ReLU(inplace=True)
self.conv1 = nn.Conv2d(1024,
128,
kernel_size=1,
stride=1,
padding=0,
bias=True)
def __init__(self, num_classes, loss={'softmax'}, aligned=False, **kwargs):
super(InceptionV4ReID, self).__init__()
self.loss = loss
base = inceptionv4()
self.features = base.features
self.classifier = nn.Linear(1536, num_classes)
self.feat_dim = 1536 # feature dimension
self.aligned = aligned
self.horizon_pool = HorizontalMaxPool2d()
if self.aligned:
self.bn = nn.BatchNorm2d(1536)
self.relu = nn.ReLU(inplace=True)
self.conv1 = nn.Conv2d(1536,
128,
kernel_size=1,
stride=1,
padding=0,
bias=True)