Page MenuHomec4science
Files F72636220

networks.py
No OneTemporary
Actions

File Metadata

Created
Tue, Jul 16, 07:50

networks.py
View Options

import torch.nn as nn
import torch.nn.functional as F
import matplotlib.pyplot as plt # plotting library
import torch
class Encoder(nn.Module):
def __init__(self,dropout):
super(Encoder, self).__init__()
self.convnet = nn.Sequential(nn.Conv2d(in_channels=3, out_channels=4, kernel_size=3,stride=2),
nn.BatchNorm2d(4),
nn.ReLU(),
nn.Dropout(dropout),
nn.Conv2d(in_channels=4, out_channels=8, kernel_size=3,stride=2),
nn.BatchNorm2d(8),
nn.ReLU(),
nn.Dropout(dropout),
nn.Conv2d(in_channels=8, out_channels=16, kernel_size=3,stride=2),
nn.BatchNorm2d(16),
nn.ReLU(),
nn.Dropout(dropout),
nn.Conv2d(in_channels=16, out_channels=32,kernel_size=3,stride=2),
nn.BatchNorm2d(32),
nn.ReLU(),
nn.Dropout(dropout),
nn.Conv2d(in_channels=32, out_channels=64, kernel_size=3,stride=2),
nn.BatchNorm2d(64),
nn.ReLU(),
nn.Dropout(dropout),
)
self.fc = nn.Sequential(nn.Linear(64 *14 * 9, 1024),
nn.ReLU(),
nn.Dropout(dropout),
nn.Linear(1024,512),
)
def forward(self, x):
output = self.convnet(x)
output = output.view(-1, 64 *14 * 9)
output = self.fc(output)
return output
class Decoder(nn.Module):
def __init__(self, dropout):
super(Decoder, self).__init__()
### Linear section
self.decoder_lin = nn.Sequential(
# First linear layer
nn.Linear(512, 1024),
nn.ReLU(),
nn.Linear(1024, 64 *14 * 9),
nn.ReLU(),
)
### Unflatten
self.unflatten = nn.Sequential(nn.Unflatten(dim=1, unflattened_size=(64, 14, 9)),
#Print
nn.BatchNorm2d(64))
### Convolutional section
self.decoder_conv = nn.Sequential(
# First transposed convolution
nn.ConvTranspose2d(64, 32, 3, stride=2, output_padding=0),
nn.ReLU(),
nn.Dropout(dropout),
nn.BatchNorm2d(32),
# Second transposed convolution
nn.ConvTranspose2d(32, 16, 3, stride=2, output_padding=0),
nn.ReLU(),
nn.Dropout(dropout),
nn.BatchNorm2d(16),
# Third transposed convolution
nn.ConvTranspose2d(16, 8, 3, stride=2, output_padding=0),
nn.ReLU(),
nn.Dropout(dropout),
nn.BatchNorm2d(8),
# fourth transposed convolution
nn.ConvTranspose2d(8, 4, 3, stride=2, output_padding=0),
nn.ReLU(),
nn.Dropout(dropout),
nn.BatchNorm2d(4),
# fivth transposed convolution
nn.ConvTranspose2d(4, 3, 3, stride=2, output_padding=1),
nn.ReLU(),
nn.Dropout(dropout),
)
def forward(self, x):
# Apply linear layers
x = self.decoder_lin(x)
x = self.unflatten(x)
x = self.decoder_conv(x)
# Apply a sigmoid to force the output to be between 0 and 1 (valid pixel values)
return x
#%%
class Discriminator(nn.Module):
def __init__(self,dropout):
super(Discriminator, self).__init__()
self.convnet = nn.ModuleList([nn.Conv2d(in_channels=3, out_channels=4, kernel_size=3,stride=2),
nn.BatchNorm2d(4),
nn.ReLU(),
nn.Dropout(dropout),
nn.Conv2d(in_channels=4, out_channels=8, kernel_size=3,stride=2),
nn.BatchNorm2d(8),
nn.ReLU(),
nn.Dropout(dropout),
nn.Conv2d(in_channels=8, out_channels=16, kernel_size=3,stride=2),
nn.BatchNorm2d(16),
nn.ReLU(),
#Print
# nn.MaxPool2d(2,2),
nn.Dropout(dropout),
nn.Conv2d(in_channels=16, out_channels=32,kernel_size=3,stride=2),
nn.BatchNorm2d(32),
nn.ReLU(),
nn.Dropout(dropout),
nn.Conv2d(in_channels=32, out_channels=64, kernel_size=3,stride=2),
nn.BatchNorm2d(64),
nn.ReLU(),
nn.Dropout(dropout),
#Print
])
self.fc = nn.ModuleList([
nn.Flatten(),
nn.Linear(8064, 512),
nn.ReLU(),
nn.Dropout(dropout),
nn.Linear(512,64),
nn.Dropout(dropout),
nn.Linear(64,1),
nn.BatchNorm1d(1),
nn.Sigmoid()]
)
def forward(self, x):
featurebank = []
for idx_l, layer in enumerate(self.convnet):
x = layer(x)
if("torch.nn.modules.activation" in str(type(layer))):
featurebank.append(x)
convout = x
for idx_l, layer in enumerate(self.fc):
x = layer(x)
if("torch.nn.modules.activation" in str(type(layer))):
featurebank.append(x)
disc_score = x
return disc_score, featurebank

Event Timeline

c4science · Help