Problem in U-NET

1

Hi,I made this U-net code

class DoubleConv(nn.Module):
    def __init__(self, in_ch, out_ch):
        super(DoubleConv, self).__init__()
        self.conv = nn.Sequential(
            nn.Conv2d(in_ch, out_ch, 1, padding=1),
            nn.BatchNorm2d(out_ch),
            nn.ReLU(inplace=True),
            nn.Conv2d(out_ch, out_ch, 3, padding=1),
            nn.BatchNorm2d(out_ch),
            nn.ReLU(inplace=True))
    def forward(self, x):
        x = self.conv(x)
        return x
class InConv(nn.Module):
    def __init__(self, in_ch, out_ch):
        super(InConv, self).__init__()
        self.conv = DoubleConv(in_ch, out_ch)
    def forward(self, x):
        x = self.conv(x)
        return x
class Down(nn.Module):
    def __init__(self, in_ch, out_ch):
        super(Down, self).__init__()
        self.mpconv = nn.Sequential(
            nn.MaxPool2d(2),
            DoubleConv(in_ch, out_ch)
        )
    def forward(self, x):
        x = self.mpconv(x)
        return x
class Up(nn.Module):
    def __init__(self, in_ch, out_ch, bilinear=True):
        super(Up, self).__init__()
        if bilinear:
            self.up = nn.Upsample(scale_factor=2, mode='bilinear', align_corners=True)
        else:
            self.up = nn.ConvTranspose2d(in_ch // 2, in_ch // 2, 2, stride=2)
        self.conv = DoubleConv(in_ch, out_ch)
    def forward(self, x1, x2):
        x1 = self.up(x1)
        diffY = x2.size()[2] - x1.size()[2]
        diffX = x2.size()[3] - x1.size()[3]
        x1 = F.pad(x1, (diffX // 2, diffX - diffX // 2,
                        diffY // 2, diffY - diffY // 2))
        x = torch.cat([x2, x1], dim=1)
        x = self.conv(x)
        return x
class OutConv(nn.Module):
    def __init__(self, in_ch, out_ch):
        super(OutConv, self).__init__()
        self.conv = nn.Conv2d(in_ch, out_ch, 1)
    def forward(self, x):
        x = self.conv(x)
        return x
class Unet(nn.Module):
    def __init__(self, in_channels, classes):
        super(Unet, self).__init__()
        self.n_channels = 3
        self.n_classes = 2
        self.inc = InConv(in_channels, 64)
        self.down1 = Down(64, 128)
        self.down2 = Down(128, 256)
        self.down3 = Down(256, 512)
        self.down4 = Down(512, 512)
        self.up1 = Up(1024, 256)
        self.up2 = Up(512, 128)
        self.up3 = Up(256, 64)
        self.up4 = Up(128, 64)
        self.outc = OutConv(64, classes)
    def forward(self, x):
        x1 = self.inc(x)
        x2 = self.down1(x1)
        x3 = self.down2(x2)
        x4 = self.down3(x3)
        x5 = self.down4(x4)
        x = self.up1(x5, x4)
        x = self.up2(x, x3)
        x = self.up3(x, x2)
        x = self.up4(x, x1)
        x = self.outc(x)
        return x

problem : RuntimeError: Given groups=1, weight of size [64, 4, 1, 1], expected input[6451, 1, 5, 5] to have 4 channels, but got 1 channels instead

help me plzz

2

It looks like there is an n_channels member of the Unet class that is defined but never used, and in_channels is defined when the Unet is created (this value seems like it is set to 4). However, the input has only a single channel? I would check how these parameters are used and whether the input data is correct because the shape seems very strange.

3

Hi @eqy yes a single channel. I prefer to get this architecture

image
image569×532 28.5 KB

help me plzz eqy

4

help me plzz :pensive: :pensive: :pensive: :pensive:

5

There is something wrong with your input. What is the input shape?

6

Hi @Hmrishav_Bandyopadhy I solved my problem, but I couldn’t do this architecture

can you help me plzz

image
image569×532 28.5 KB

7

This looks like the Unet architecture with some blocks skipped. You can see an example implementation in this tutorial video here

8

Hi @Hmrishav_Bandyopadhy I tested this code but

problem : RuntimeError: only batches of spatial targets supported (3D tensors) but got targets of dimension: 1

I prefer to detect two classes

9

Once again if you have targets of dimension 1 then there is an issue with the data that you are using. UNets are segmentation models which means that the input and output would be 2D data…Add a dimension for batch size and you have 3D tensors

10

just segmentation unet? do not make a classification

11

When I tested this code

class UNet(nn.Module):
  def __init__(self): 
   super(UNet, self).__init__()
   self.max_pool = nn.MaxPool2d(2,2)
   self.down_conv1 = doubleconv(1,64)
   self.down_conv2= doubleconv(64,128)
   self.down_conv3= doubleconv(128,256)
   self.down_conv4= doubleconv(256,512)
   self.down_conv5= doubleconv(512,1024)

  def forward (self,x) : 

     x1= self.down_conv1(x)
     #x2= self.max_pool(x1)
     x3= self.down_conv2(x1)
     #x4= self.max_pool(x3)
     x5= self.down_conv3(x3)
     #x6= self.max_pool(x5)
     x7= self.down_conv4(x5)
     #x8= self.max_pool(x7)
     x9= self.down_conv5(x7)

RuntimeError: Given groups=1, weight of size [64, 1, 3, 3], expected input[1481, 128, 1, 1] to have 1 channels, but got 128 channels instead

help me plzz

12

I solved my problem :blush: :blush: :blush: