MCPcopy
hub / github.com/thuml/Transfer-Learning-Library / ReidResNet

Class ReidResNet

tllib/vision/models/reid/resnet.py:10–33  ·  view source on GitHub ↗

r"""Modified `ResNet` architecture for ReID from `Mutual Mean-Teaching: Pseudo Label Refinery for Unsupervised Domain Adaptation on Person Re-identification (ICLR 2020) `_. We change stride of :math:`layer4\_group1\_conv2, layer4\_group1\_downsample1` to

Source from the content-addressed store, hash-verified

8
9
10class ReidResNet(ResNet):
11 r"""Modified `ResNet` architecture for ReID from `Mutual Mean-Teaching: Pseudo Label Refinery for Unsupervised
12 Domain Adaptation on Person Re-identification (ICLR 2020) <https://arxiv.org/pdf/2001.01526.pdf>`_. We change stride
13 of :math:`layer4\_group1\_conv2, layer4\_group1\_downsample1` to 1. During forward pass, we will not activate
14 `self.relu`. Please refer to source code for details.
15 """
16
17 def __init__(self, *args, **kwargs):
18 super(ReidResNet, self).__init__(*args, **kwargs)
19 self.layer4[0].conv2.stride = (1, 1)
20 self.layer4[0].downsample[0].stride = (1, 1)
21
22 def forward(self, x):
23 x = self.conv1(x)
24 x = self.bn1(x)
25 # x = self.relu(x)
26 x = self.maxpool(x)
27
28 x = self.layer1(x)
29 x = self.layer2(x)
30 x = self.layer3(x)
31 x = self.layer4(x)
32
33 return x
34
35
36def _reid_resnet(arch, block, layers, pretrained, progress, **kwargs):

Callers 1

_reid_resnetFunction · 0.85

Calls

no outgoing calls

Tested by

no test coverage detected