mnist_archs.py

import torch
import torch.nn as nn
import torch.nn.functional as F
# F.Chollet architecture
class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 32, kernel_size=3)
        self.conv2 = nn.Conv2d(32, 64, kernel_size=3)
        # dropout 0.25
        self.fc1 = nn.Linear(1600, 128)
        # dropout 0.5
        self.fc2 = nn.Linear(128, 10)

    def forward(self, x):
        x = F.relu(F.max_pool2d( self.conv1(x), 2 ) )
        x = F.relu(F.max_pool2d( self.conv2(x), 2 ) )      
        x = x.view(-1, 1600)
        x = F.relu(self.fc1(x))
        x = self.fc2(x)
        return F.log_softmax(x, dim=1)
    
    def extract(self, x,verbose=False):
        out1 = F.relu(F.max_pool2d(self.conv1(x),    2 ) )
        out2 = F.relu(F.max_pool2d(self.conv2(out1), 2 ) )       
        t = out2.view(-1, 1600)
        out3 = F.relu(self.fc1(t))
        t = self.fc2(out3)
        out4 = F.log_softmax(t, dim=1)
        
        if verbose == True:
            print(out1.size())
            print(out2.size())
            print(out3.size())
            print(out4.size())
        
        return out1, out2, out3, out4
    
    def extract_all(self, x,verbose=False):
        out1 = self.conv1(x)
        out2 = F.relu(F.max_pool2d(out1,2))
        out3 = self.conv2(out2)
        out4 = F.relu(F.max_pool2d(out3,2))  
        t = out4.view(-1, 1600)
        out5 = F.relu(self.fc1(t))
        t = self.fc2(out5)
        out6 = F.log_softmax(t, dim=1)
        
        if verbose == True:
            print(out1.size())
            print(out2.size())
            print(out3.size())
            print(out4.size())
            print(out5.size())
            print(out6.size())
        return out1, out2, out3, out4, out5, out6