Resnet152

[ChrisXULC]

请问如果我想把网络换成Resnet152，
class ResNet(nn.Module):
def init(self,c_in):
super(ResNet,self).init()
self.block0 = nn.Sequential(nn.Conv2d(c_in, 32, 3, 1, 1),nn.BatchNorm2d(32, momentum=0.01))
self.block1 = self._make_layer(32, 32, 2, 3)
self.block2 = self._make_layer(32, 64, 2, 8)
self.block3 = self._make_layer(64, 128, (2,1),8)
self.block4 = self._make_layer(128, 256, (2,1),36)
self.block5 = self._make_layer(256, 512, (2,1), 3)
然后进行模型训练 loss不断减小 但是acc一直为0 是为什么呀？十分谢谢。

[Canjie-Luo]

请问有更多细节吗，否则我不好回复～