我对DL（以及一般而言的编码）非常陌生。我正在pytorch中松散地基于V-Net实现一个非常简单的3D CNN（下面的代码）。目前，我可以使代码在小型测试图像（形状：（1,1,128,128,32））上运行，但是例如，如果我输入更大的值（1,1,256,256,128），该模型将使Colab崩溃。可以预料，或者如果我在这里犯了一些明显的错误，请先谢谢。
class Test3D(nn.Module):
def __init__(self):
super(Test3D,self).__init__()
self.input_layer = self._conv_input()
self.conv_layer1 = self.ResidualBlock(32, 40)
self.conv_layer2 = self.ResidualBlock(40, 48)
self.conv_layer3 = self.ResidualBlock(48,56)
self.conv_layer4 = self.ResidualBlock(56,48)
self.conv_layer5 = self.ResidualBlock(48,40)
self.conv_layer6 = self.ResidualBlock(40,32)
self.conv_layer7 = self._conv_output()
def _conv_input(self):
conv_layer= nn.Sequential(
nn.Conv3d(1, 32, kernel_size=(3, 3, 3), stride=2,padding=1)
)
return conv_layer
def _conv_output(self):
conv_layer= nn.Sequential(
nn.ConvTranspose3d(32, 6, kernel_size=2, stride=2,padding=0)
)
return conv_layer
def ResidualBlock(self,in_c,out_c):
conv_layer=nn.Sequential(
nn.Conv3d(in_c,out_c,kernel_size=(3,3,3),padding=1),
nn.Conv3d(out_c,out_c,kernel_size=(3,3,3),padding=1),
ContBatchNorm3d(out_c),
nn.ReLU()
)
return conv_layer
def forward(self, x):
out = self.input_layer(x)
print(out.shape)
out = self.conv_layer1(out)
print(out.shape)
out = self.conv_layer2(out)
print(out.shape)
out = self.conv_layer3(out)
print(out.shape)
out = self.conv_layer4(out)
print(out.shape)
out = self.conv_layer5(out)
print(out.shape)
out = self.conv_layer6(out)
print(out.shape)
out = self.conv_layer7(out)
print(out.shape)