CNN+LSTM

CNN+LSTM

Testing MAE in epoch 0: 10.3750925064

Testing MAE in epoch 1: 10.4844141006

Testing MAE in epoch 2: 10.270113945

Testing MAE in epoch 3: 10.3185472488

Testing MAE in epoch 4: 10.2815971375

Testing MAE in epoch 5: 10.6698894501

Testing MAE in epoch 6: 10.1276693344

Testing MAE in epoch 7: 10.0981492996

Testing MAE in epoch 8: 10.1957941055

Testing MAE in epoch 9: 10.5166063309

Testing MAE in epoch 10: 10.5095157623

Testing MAE in epoch 11: 10.3940849304

Testing MAE in epoch 12: 10.305264473

Testing MAE in epoch 13: 10.298617363

Testing MAE in epoch 14: 10.823056221

Testing MAE in epoch 15: 10.0867042542

Testing MAE in epoch 16: 10.3575077057

Testing MAE in epoch 17: 10.3023538589

Testing MAE in epoch 18: 10.313621521

Testing MAE in epoch 19: 10.1169481277 


class ConvNet(nn.Module):
    def __init__(self, num_classes=1):
        super(ConvNet, self).__init__()
        self.layer1 = nn.Sequential(
            nn.Conv2d(1, 16,kernel_size=2, stride=1, padding=2),
            nn.BatchNorm2d(16),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, stride=2))
        self.layer2 = nn.Sequential(
            nn.Conv2d(16, 32, kernel_size=2, stride=1, padding=2),
            nn.BatchNorm2d(32),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, stride=2))
        self.lstm = nn.LSTM(16128,200)
        self.fc = nn.Linear(200, 1)
        
    def forward(self, x):
        out = self.layer1(x)
        #print(out.shape)
        out = self.layer2(out)


        
        out = out.reshape(out.size(0), -1)
        #print(out.shape)
        out = out.unsqueeze(0)
        out, hid = self.lstm(out)
        #print(out.shape)
        out = out.squeeze(0)
        out = self.fc(out)
        #print(out.shape)
        return out


class ConvNet(nn.Module):
    def __init__(self, num_classes=1):
        super(ConvNet, self).__init__()
        self.layer1 = nn.Sequential(
            nn.Conv2d(1, 16,kernel_size=2, stride=1, padding=2),
            nn.BatchNorm2d(16),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, stride=2))
        self.layer2 = nn.Sequential(
            nn.Conv2d(16, 32, kernel_size=2, stride=1, padding=2),
            nn.BatchNorm2d(32),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, stride=2))
        self.layer3 = nn.Sequential(
            nn.Conv2d(32, 64, kernel_size=2, stride=1, padding=2),
            nn.BatchNorm2d(64),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=4, stride=2))
        self.lstm = nn.LSTM(8064,200)
        self.fc = nn.Linear(200, 1)
        
    def forward(self, x):
        out = self.layer1(x)
        #print(out.shape)
        out = self.layer2(out)
        out = self.layer3(out)


        
        out = out.reshape(out.size(0), -1)
        #print(out.shape)
        out = out.unsqueeze(0)
        out, hid = self.lstm(out)
        #print(out.shape)
        out = out.squeeze(0)
        out = self.fc(out)
        #print(out.shape)
        return out

Testing MAE in epoch 0: 10.2088651657

Testing MAE in epoch 1: 10.2906723022

Testing MAE in epoch 2: 10.181760788

Testing MAE in epoch 3: 10.5987415314

Testing MAE in epoch 4: 10.182387352

Testing MAE in epoch 5: 10.3811950684

Testing MAE in epoch 6: 10.3703622818

Testing MAE in epoch 7: 10.3837184906

Testing MAE in epoch 8: 10.2380847931

Testing MAE in epoch 9: 10.9325876236

Testing MAE in epoch 10: 10.3778715134

Testing MAE in epoch 11: 10.1646280289

Testing MAE in epoch 12: 10.2416553497

Testing MAE in epoch 13: 10.155752182

Testing MAE in epoch 14: 10.2135791779

Testing MAE in epoch 15: 10.2569732666

Testing MAE in epoch 16: 10.2894687653

Testing MAE in epoch 17: 10.7199611664

Testing MAE in epoch 18: 10.4396028519

Testing MAE in epoch 19: 10.4319229126 


class ConvNet(nn.Module):
    def __init__(self, num_classes=1):
        super(ConvNet, self).__init__()
        self.layer1 = nn.Sequential(
            nn.Conv2d(1, 16,kernel_size=2, stride=1, padding=2),
            nn.BatchNorm2d(16),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, stride=2))
        self.layer2 = nn.Sequential(
            nn.Conv2d(16, 32, kernel_size=2, stride=1, padding=2),
            nn.BatchNorm2d(32),
            nn.ReLU(),
            nn.MaxPool2d(kernel_size=2, stride=2))
        self.lstm = nn.LSTM(16128,200,3)
        self.fc = nn.Linear(200, 1)
        
    def forward(self, x):
        out = self.layer1(x)
        #print(out.shape)
        out = self.layer2(out)


        
        out = out.reshape(out.size(0), -1)
        #print(out.shape)
        out = out.unsqueeze(0)
        out, hid = self.lstm(out)
        #print(out.shape)
        out = out.squeeze(0)
        out = self.fc(out)
        #print(out.shape)
        return out

Testing MAE in epoch 0: 10.0391073227

Testing MAE in epoch 1: 10.5044641495

Testing MAE in epoch 2: 10.6294088364

Testing MAE in epoch 3: 10.1677951813

Testing MAE in epoch 4: 10.5848722458

Testing MAE in epoch 5: 10.3339357376

Testing MAE in epoch 6: 10.3954181671

Testing MAE in epoch 7: 10.3198976517

Testing MAE in epoch 8: 10.0761852264

Testing MAE in epoch 9: 10.6136617661

Testing MAE in epoch 10: 9.92497634888

Testing MAE in epoch 11: 10.2302398682

Testing MAE in epoch 12: 10.3068552017

Testing MAE in epoch 13: 10.18638134

Testing MAE in epoch 14: 10.0027036667

Testing MAE in epoch 15: 10.5765810013

Testing MAE in epoch 16: 10.223285675

Testing MAE in epoch 17: 10.4815912247

Testing MAE in epoch 18: 10.6809406281

Testing MAE in epoch 19: 10.649895668


CNN+LSTM with 1000 outputs

Testing MAE in epoch 0: 10.1571788788

Testing MAE in epoch 1: 10.0599842072

Testing MAE in epoch 2: 10.590801239

Testing MAE in epoch 3: 9.95899200439

Testing MAE in epoch 4: 9.99138736725

Testing MAE in epoch 5: 10.1518325806

Testing MAE in epoch 6: 10.4582710266

Testing MAE in epoch 7: 10.9011421204

Testing MAE in epoch 8: 10.2009010315

Testing MAE in epoch 9: 10.4495191574

Testing MAE in epoch 10: 10.2944841385

Testing MAE in epoch 11: 10.0457792282

Testing MAE in epoch 12: 10.1445274353

CNN + 5 LSTM

Testing MAE in epoch 0: 10.7628278732

Testing MAE in epoch 1: 10.2336072922

Testing MAE in epoch 2: 10.2210321426

Testing MAE in epoch 3: 10.5542860031

Testing MAE in epoch 4: 10.7996482849

Testing MAE in epoch 5: 10.1400461197

Testing MAE in epoch 6: 10.1726551056

Testing MAE in epoch 7: 11.0996789932

Testing MAE in epoch 8: 10.2891817093

Testing MAE in epoch 9: 10.8611316681

Testing MAE in epoch 10: 10.2157659531

Testing MAE in epoch 11: 10.1307249069


Depth :

  • 1 - 9.92
  • 2 - 10.10
  • 3- 10.01
  • 4- 10.00
  • 5 - 9.97

Padding = 1

10.221

More layers : 10.08