Scribd Logo
Upload

Welcome to Scribd!

  • Lab 3

    Uploaded by

    Rafał Kuśmierz
    2 views4 pages
    The document describes the architecture of several neural network models for image classification. It defines multiple convolutional neural network (CNN) models with variations in the number of convolutional and max pooling layers, as well as the number and size of fully connected layers at the end of the models. The models take in input images, pass them through convolutional and pooling layers to extract features, flatten the output, and use fully connected layers to perform classification with 50 output classes.

    Original Description:

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    The document describes the architecture of several neural network models for image classification. It defines multiple convolutional neural network (CNN) models with variations in the number of convolutional and max pooling layers, as well as the number and size of fully connected layers at the end of the models. The models take in input images, pass them through convolutional and pooling layers to extract features, flatten the output, and use fully connected layers to perform classification with 50 output classes.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    2 views4 pages

    Lab 3

    Uploaded by

    Rafał Kuśmierz
    The document describes the architecture of several neural network models for image classification. It defines multiple convolutional neural network (CNN) models with variations in the number of convolutional and max pooling layers, as well as the number and size of fully connected layers at the end of the models. The models take in input images, pass them through convolutional and pooling layers to extract features, flatten the output, and use fully connected layers to perform classification with 50 output classes.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    of 4

    Lab 3

    piątek, 18 listopada 2022 23:03

    3, 32, 32 3, 64, 64 3, 64, 64 3, 64, 64


    6, 28, 28 6, 60, 60 60, 60
    6, 14, 14 6, 30, 30 6, 60, 60 56, 56
    16, 10, 10 16, 26, 26 6, 30, 30 33, 33
    16,5,5 16, 13, 13
    12, 28, 28 32, 32
    12, 14, 14 28, 28
    24, 24
    32, 10, 10
    32, 5, 5 12, 12

    classNet(nn.Module):
    def__init__(self):
    super().__init__()
    ##Warstwakonwolucyjna
    self.conv1=nn.Conv2d(in_channels=3,out_channels=6,kernel_size=
    5,stride=1,padding=0)
    ##Warstwamaxpooling
    self.pool1=nn.MaxPool2d(kernel_size=2,stride=2)
    self.conv2=nn.Conv2d(6,16,5)
    self.pool2=nn.MaxPool2d(2)
    self.fc1=nn.Linear(16*13*13,800)
    self.fc2=nn.Linear(800,420)
    self.fc3=nn.Linear(420,50)

    defforward(self,x):
    x=self.pool1(F.relu(self.conv1(x)))
    x=self.pool2(F.relu(self.conv2(x)))
    x=torch.flatten(x,1)#flattenalldimensionsexceptbatch
    x=F.relu(self.fc1(x))
    x=F.relu(self.fc2(x))
    x=self.fc3(x)
    Returnx

    classNet(nn.Module):
    def__init__(self):
    super().__init__()
    ##Warstwakonwolucyjna
    self.conv0=nn.Conv2d(in_channels=3,out_channels=6,kernel_size=
    5,stride=1,padding=0)
    self.pool0=nn.MaxPool2d(kernel_size=2,stride=2)

    self.conv1=nn.Conv2d(in_channels=6,out_channels=12,kernel_size=
    3,stride=1,padding=0)
    ##Warstwamaxpooling
    self.pool1=nn.MaxPool2d(kernel_size=2,stride=2)
    self.conv2=nn.Conv2d(12,32,5)
    self.pool2=nn.MaxPool2d(2)
    self.fc1=nn.Linear(32*5*5,240)
    self.fc2=nn.Linear(240,168)
    self.fc3=nn.Linear(168,50)

    defforward(self,x):
    x=self.pool0(F.relu(self.conv0(x)))
    x=self.pool1(F.relu(self.conv1(x)))
    x=self.pool2(F.relu(self.conv2(x)))
    x=torch.flatten(x,1)#flattenalldimensionsexceptbatch
    x=F.relu(self.fc1(x))
    x=F.relu(self.fc2(x))
    x=self.fc3(x)
    Returnx

    SSNE Strona 1
    Returnx

    classNet(nn.Module):
    def__init__(self):
    super().__init__()
    ##Warstwakonwolucyjna
    self.conv0=nn.Conv2d(in_channels=3,out_channels=6,kernel_size=5,stride=1,padding=0)
    self.pool0=nn.MaxPool2d(kernel_size=2,stride=2)

    self.conv1=nn.Conv2d(in_channels=6,out_channels=12,kernel_size=3,stride=1,padding=0)
    ##Warstwamaxpooling
    self.pool1=nn.MaxPool2d(kernel_size=2,stride=2)
    self.conv2=nn.Conv2d(12,32,5)
    self.pool2=nn.MaxPool2d(2,stride=1)
    self.fc1=nn.Linear(32*9*9,600)#chcemy2000->600
    self.fc2=nn.Linear(600,420)
    self.fc3=nn.Linear(420,50)

    defforward(self,x):
    x=self.pool0(F.relu(self.conv0(x)))
    x=self.pool1(F.relu(self.conv1(x)))
    x=self.pool2(F.relu(self.conv2(x)))
    x=torch.flatten(x,1)#flattenalldimensionsexceptbatch
    x=F.relu(self.fc1(x))
    x=F.relu(self.fc2(x))
    x=self.fc3(x)
    Returnx

    3, 64, 64

    1
    6, 60, 60
    6, 58, 58

    2
    12, 54, 54
    12, 52, 52

    3
    18, 48, 48
    18, 46, 46

    4
    24, 42, 42
    24, 40, 40

    5
    24, 36, 36
    24, 18, 18

    6
    36, 14, 14
    36, 7, 7

    SSNE Strona 2
    36, 7, 7

    classNet(nn.Module):
    def__init__(self):
    super().__init__()
    ##Warstwakonwolucyjna

    self.conv0=nn.Conv2d(in_channels=
    3,out_channels=6,kernel_size=
    5,stride=1,padding=0)
    self.pool0
    =nn.MaxPool2d(kernel_size=
    2,stride=2)
    self.conv1=nn.Conv2d(in_channels=
    6,out_channels=12,kernel_size=
    5,stride=1,padding=0)
    self.pool1
    =nn.MaxPool2d(kernel_size=
    2,stride=2)
    self.conv2=nn.Conv2d(12,16,4)
    self.pool2=nn.MaxPool2d(2)
    self.fc1=nn.Linear(16*5*5,120)
    self.fc2=nn.Linear(120,84)
    self.fc3=nn.Linear(84,50)

    defforward(self,x):
    x=self.pool0(F.relu(self.conv0(x)))
    x=self.pool1(F.relu(self.conv1(x)))
    x=self.pool2(F.relu(self.conv2(x)))
    x=torch.flatten(x,1)#flattenalldimensio
    nsexceptbatch
    x=F.relu(self.fc1(x))
    x=F.relu(self.fc2(x))
    x=self.fc3(x)
    returnx

    classNet(nn.Module):
    def__init__(self):
    super().__init__()
    ##Warstwakonwolucyjna
    self.conv1=nn.Conv2d(in_channels=
    3,out_channels=6,kernel_size=5,stride=1,padding=
    0)
    ##Warstwamaxpooling
    self.pool1=nn.MaxPool2d(kernel_size=2,stride=2)
    self.conv2=nn.Conv2d(6,16,5)
    self.pool2=nn.MaxPool2d(2)
    self.fc1=nn.Linear(16*13*13,800)
    self.fc2=nn.Linear(800,600)
    self.fc3=nn.Linear(600,420)
    self.fc4=nn.Linear(420,50)

    defforward(self,x):
    x=self.pool1(F.relu(self.conv1(x)))
    x=self.pool2(F.relu(self.conv2(x)))
    x=torch.flatten(x,1)#flattenalldimensionsexceptbatch
    x=F.relu(self.fc1(x))
    x=F.relu(self.fc2(x))
    x=F.relu(self.fc3(x))
    x=self.fc4(x)
    returnx

    SSNE Strona 3
    classNet(nn.Module):
    def__init__(self):
    super().__init__()
    ##Warstwakonwolucyjna
    self.conv1 Tutaj gorzej niż wersja poprzednio
    =nn.Conv2d(in_channels=
    3,out_channels=6,kernel_size=
    5,stride=1,padding=0)
    ##Warstwamaxpooling
    self.pool1
    =nn.MaxPool2d(kernel_size=
    2,stride=2)
    self.bn1=nn.BatchNorm2d(6)
    self.conv2=nn.Conv2d(6,16,5)
    self.pool2=nn.MaxPool2d(2)
    self.bn2=nn.BatchNorm2d(16)
    self.fc1=nn.Linear(16*13*13,800)
    self.fc2=nn.Linear(800,600)
    self.fc3=nn.Linear(600,420)
    self.fc4=nn.Linear(420,50)

    defforward(self,x):
    x=self.pool1(F.relu(self.conv1(x)))
    x=self.bn1(x)
    x=self.pool2(F.relu(self.conv2(x)))
    x=self.bn2(x)
    x=torch.flatten(x,1)#flattenalldimen
    sionsexceptbatch
    x=F.relu(self.fc1(x))
    x=F.relu(self.fc2(x))
    x=F.relu(self.fc3(x))
    x=self.fc4(x)
    returnx

    SSNE Strona 4

    You might also like