train-model.py

Train the model using several architectures

# Load the model (Download only at first time loading the model)
model_resnet = models.resnet18(pretrained=True)
model_vgg16 = models.vgg16(pretrained=True)
model_alexnet = models.alexnet(pretrained=True)

# Define the function to train the model
def train_model(model, criterion, optimizer, scheduler, num_epochs=25):
    since = time.time()
    
    best_model_wts = copy.deepcopy(model.state_dict())
    best_acc = 0.0
    
    for epoch in range(num_epochs):
        print('Epoch {}/{}'.format(epoch, num_epochs - 1))
        print('-' * 10)
        
        for phase in ['train', 'val']:
            if phase == 'train':
                model.train()
            else:
                model.eval()
                
            running_loss = 0.0
            running_corrects = 0
            
            for inputs, labels in dataloaders[phase]:
                inputs = inputs.to(device)
                labels = labels.to(device)
                
                optimizer.zero_grad()
                
                with torch.set_grad_enabled(phase == 'train'):
                    outputs = model(inputs)
                    _, preds = torch.max(outputs, 1)
                    loss = criterion(outputs, labels)
                    
                    if phase == 'train':
                        loss.backward()
                        optimizer.step()
                        
                running_loss += loss.item() * inputs.size(0)
                running_corrects += torch.sum(preds == labels.data)
                
            if phase == 'train':
                scheduler.step()
                
            epoch_loss = running_loss / data_size[phase]
            epoch_acc = running_corrects.double() / data_size[phase]
            
            print('{} Loss: {:.4f} Acc: {:.4f}'.format(
                phase, epoch_loss, epoch_acc))
            
            if phase == 'val' and epoch_acc > best_acc:
                best_acc = epoch_acc
                best_model_wts = copy.deepcopy(model.state_dict())
        print()
    
    time_elapsed = time.time() - since
    print('Training complete in {:.0f}m {:.0f}s'.format(
        time_elapsed // 60, time_elapsed % 60))
    print('Best val Acc: {:4f}'.format(best_acc))
    
    model.load_state_dict(best_model_wts)
    return model

# Set the requires_grad on each parameter to false, 
# so it will not calculate the gradients
for param in model_vgg16.parameters():
    param.requires_grad = False

for param in model_resnet.parameters():
    param.requires_grad = False

for param in model_alexnet.parameters():
    param.requires_grad = False

# Set the new final layer for our new dataset
num_ftrs = model_vgg16.classifier[6].in_features
model_vgg16.classifier[6] = nn.Linear(num_ftrs, len(class_names))

num_ftrs = model_resnet.fc.in_features
model_resnet.fc = nn.Linear(num_ftrs, len(class_names))

num_ftrs = model_alexnet.classifier[6].in_features
model_alexnet.classifier[6] = nn.Linear(num_ftrs, len(class_names))

# Enable GPU for the model
model_vgg16 = model_vgg16.to(device)
model_resnet = model_resnet.to(device)
model_alexnet = model_alexnet.to(device)

# Set the loss function
criterion = nn.CrossEntropyLoss()

# Set the optimizer and the scheduler to update the weights, and train the model.
optimizer_conv = optim.SGD(model_vgg16.classifier[6].parameters(), lr=0.001, momentum=0.9)
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_conv, step_size=7, gamma=0.1)
model_vgg16 = train_model(model_vgg16, criterion, optimizer_conv, exp_lr_scheduler, num_epochs=25)

optimizer_conv = optim.SGD(model_resnet.fc.parameters(), lr=0.001, momentum=0.9)
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_conv, step_size=7, gamma=0.1)
model_resnet = train_model(model_resnet, criterion, optimizer_conv, exp_lr_scheduler, num_epochs=25)

optimizer_conv = optim.SGD(model_alexnet.classifier[6].parameters(), lr=0.001, momentum=0.9)
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_conv, step_size=7, gamma=0.1)
model_alexnet = train_model(model_alexnet, criterion, optimizer_conv, exp_lr_scheduler, num_epochs=25)