fedex/mt
Public Access
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

initial work for elpv and subter datasets

Browse Source 
elpv as example dataset/implementation
subter with final dataset
This commit is contained in:
Jan Kowalczyk
2024-06-28 11:40:19 +02:00
parent 71f9662022
commit d6a019a8bb
13 changed files with 1585 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

74
Deep-SAD-PyTorch/src/networks/elpv_LeNet.py Normal file
View File

@@ -0,0 +1,74 @@
import torch
import torch.nn as nn
import torch.nn.functional as F
from base.base_net import BaseNet
class ELPV_LeNet(BaseNet):
def __init__(self, rep_dim=256):
super().__init__()
self.rep_dim = rep_dim
self.pool = nn.MaxPool2d(2, 2)
self.conv1 = nn.Conv2d(1, 8, 5, bias=False, padding=2)
self.bn1 = nn.BatchNorm2d(8, eps=1e-04, affine=False)
self.conv2 = nn.Conv2d(8, 4, 5, bias=False, padding=2)
self.bn2 = nn.BatchNorm2d(4, eps=1e-04, affine=False)
self.fc1 = nn.Linear(4 * 75 * 75, self.rep_dim, bias=False)
def forward(self, x):
x = x.view(-1, 1, 300, 300)
x = self.conv1(x)
x = self.pool(F.leaky_relu(self.bn1(x)))
x = self.conv2(x)
x = self.pool(F.leaky_relu(self.bn2(x)))
x = x.view(int(x.size(0)), -1)
x = self.fc1(x)
return x
class ELPV_LeNet_Decoder(BaseNet):
def __init__(self, rep_dim=256):
super().__init__()
self.rep_dim = rep_dim
# Decoder network
self.fc3 = nn.Linear(self.rep_dim, 2888, bias=False)
self.bn1d2 = nn.BatchNorm1d(2888, eps=1e-04, affine=False)
self.deconv1 = nn.ConvTranspose2d(2, 4, 5, bias=False, padding=2)
self.bn3 = nn.BatchNorm2d(4, eps=1e-04, affine=False)
self.deconv2 = nn.ConvTranspose2d(4, 8, 5, bias=False, padding=3)
self.bn4 = nn.BatchNorm2d(8, eps=1e-04, affine=False)
self.deconv3 = nn.ConvTranspose2d(8, 1, 5, bias=False, padding=2)
def forward(self, x):
x = self.bn1d2(self.fc3(x))
x = x.view(int(x.size(0)), 2, 38, 38)
x = F.interpolate(F.leaky_relu(x), scale_factor=2)
x = self.deconv1(x)
x = F.interpolate(F.leaky_relu(self.bn3(x)), scale_factor=2)
x = self.deconv2(x)
x = F.interpolate(F.leaky_relu(self.bn4(x)), scale_factor=2)
x = self.deconv3(x)
x = torch.sigmoid(x)
return x
class ELPV_LeNet_Autoencoder(BaseNet):
def __init__(self, rep_dim=256):
super().__init__()
self.rep_dim = rep_dim
self.encoder = ELPV_LeNet(rep_dim=rep_dim)
self.decoder = ELPV_LeNet_Decoder(rep_dim=rep_dim)
def forward(self, x):
x = self.encoder(x)
x = self.decoder(x)
return x

18
Deep-SAD-PyTorch/src/networks/main.py
View File

@@ -1,4 +1,6 @@
from .mnist_LeNet import MNIST_LeNet, MNIST_LeNet_Autoencoder
from .elpv_LeNet import ELPV_LeNet, ELPV_LeNet_Autoencoder
from .subter_LeNet import SubTer_LeNet, SubTer_LeNet_Autoencoder
from .fmnist_LeNet import FashionMNIST_LeNet, FashionMNIST_LeNet_Autoencoder
from .cifar10_LeNet import CIFAR10_LeNet, CIFAR10_LeNet_Autoencoder
from .mlp import MLP, MLP_Autoencoder
@@ -11,6 +13,8 @@ def build_network(net_name, ae_net=None):
implemented_networks = (
"mnist_LeNet",
"elpv_LeNet",
"subter_LeNet",
"mnist_DGM_M2",
"mnist_DGM_M1M2",
"fmnist_LeNet",
@@ -39,6 +43,12 @@ def build_network(net_name, ae_net=None):
if net_name == "mnist_LeNet":
net = MNIST_LeNet()
if net_name == "subter_LeNet":
net = SubTer_LeNet()
if net_name == "elpv_LeNet":
net = ELPV_LeNet()
if net_name == "mnist_DGM_M2":
net = DeepGenerativeModel(
[1 * 28 * 28, 2, 32, [128, 64]], classifier_net=MNIST_LeNet
@@ -118,6 +128,8 @@ def build_autoencoder(net_name):
"""Builds the corresponding autoencoder network."""
implemented_networks = (
"elpv_LeNet",
"subter_LeNet",
"mnist_LeNet",
"mnist_DGM_M1M2",
"fmnist_LeNet",
@@ -139,6 +151,12 @@ def build_autoencoder(net_name):
if net_name == "mnist_LeNet":
ae_net = MNIST_LeNet_Autoencoder()
if net_name == "subter_LeNet":
ae_net = SubTer_LeNet_Autoencoder()
if net_name == "elpv_LeNet":
ae_net = ELPV_LeNet_Autoencoder()
if net_name == "mnist_DGM_M1M2":
ae_net = VariationalAutoencoder([1 * 28 * 28, 32, [128, 64]])

70
Deep-SAD-PyTorch/src/networks/subter_LeNet.py Normal file
View File

@@ -0,0 +1,70 @@
import torch
import torch.nn as nn
import torch.nn.functional as F
from base.base_net import BaseNet
class SubTer_LeNet(BaseNet):
def __init__(self, rep_dim=1024):
super().__init__()
self.rep_dim = rep_dim
self.pool = nn.MaxPool2d(2, 2)
self.conv1 = nn.Conv2d(1, 8, 5, bias=False, padding=2)
self.bn1 = nn.BatchNorm2d(8, eps=1e-04, affine=False)
self.conv2 = nn.Conv2d(8, 4, 5, bias=False, padding=2)
self.bn2 = nn.BatchNorm2d(4, eps=1e-04, affine=False)
self.fc1 = nn.Linear(4 * 512 * 8, self.rep_dim, bias=False)
def forward(self, x):
x = x.view(-1, 1, 32, 2048)
x = self.conv1(x)
x = self.pool(F.leaky_relu(self.bn1(x)))
x = self.conv2(x)
x = self.pool(F.leaky_relu(self.bn2(x)))
x = x.view(int(x.size(0)), -1)
x = self.fc1(x)
return x
class SubTer_LeNet_Decoder(BaseNet):
def __init__(self, rep_dim=1024):
super().__init__()
self.rep_dim = rep_dim
# Decoder network
self.fc3 = nn.Linear(self.rep_dim, 4 * 512 * 8, bias=False)
self.bn3 = nn.BatchNorm2d(4, eps=1e-04, affine=False)
self.deconv1 = nn.ConvTranspose2d(4, 8, 5, bias=False, padding=2)
self.bn4 = nn.BatchNorm2d(8, eps=1e-04, affine=False)
self.deconv2 = nn.ConvTranspose2d(8, 1, 5, bias=False, padding=2)
def forward(self, x):
x = self.fc3(x)
x = x.view(int(x.size(0)), 4, 8, 512)
x = F.interpolate(F.leaky_relu(self.bn3(x)), scale_factor=2)
x = self.deconv1(x)
x = F.interpolate(F.leaky_relu(self.bn4(x)), scale_factor=2)
x = self.deconv2(x)
x = torch.sigmoid(x)
return x
class SubTer_LeNet_Autoencoder(BaseNet):
def __init__(self, rep_dim=1024):
super().__init__()
self.rep_dim = rep_dim
self.encoder = SubTer_LeNet(rep_dim=rep_dim)
self.decoder = SubTer_LeNet_Decoder(rep_dim=rep_dim)
def forward(self, x):
x = self.encoder(x)
x = self.decoder(x)
return x