Transfer Learning Tutorial

Author: Sasank Chilamkurthy

In this tutorial, you will learn how to train your network using transfer learning. You can read more about the transfer learning at cs231n notes

Quoting these notes,

In practice, very few people train an entire Convolutional Network from scratch (with random initialization), because it is relatively rare to have a dataset of sufficient size. Instead, it is common to pretrain a ConvNet on a very large dataset (e.g. ImageNet, which contains 1.2 million images with 1000 categories), and then use the ConvNet either as an initialization or a fixed feature extractor for the task of interest.

These two major transfer learning scenarios look as follows:

• Finetuning the convnet: Instead of random initializaion, we initialize the network with a pretrained network, like the one that is trained on imagenet 1000 dataset. Rest of the training looks as usual.
• ConvNet as fixed feature extractor: Here, we will freeze the weights for all of the network except that of the final fully connected layer. This last fully connected layer is replaced with a new one with random weights and only this layer is trained.

# License: BSD
# Author: Sasank Chilamkurthy

from __future__ import print_function, division

import torch
import torch.nn as nn
import torch.optim as optim
from torch.optim import lr_scheduler
import numpy as np
import torchvision
from torchvision import datasets, models, transforms
import matplotlib.pyplot as plt
import time
import os
import copy

plt.ion()   # interactive mode

Load Data

We will use torchvision and torch.utils.data packages for loading the data.

The problem we’re going to solve today is to train a model to classify ants and bees. We have about 120 training images each for ants and bees. There are 75 validation images for each class. Usually, this is a very small dataset to generalize upon, if trained from scratch. Since we are using transfer learning, we should be able to generalize reasonably well.

This dataset is a very small subset of imagenet.

Note

Download the data from here and extract it to the current directory.

# Data augmentation and normalization for training
# Just normalization for validation
data_transforms = {
    'train': transforms.Compose([
        transforms.RandomResizedCrop(224),
        transforms.RandomHorizontalFlip(),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
    'val': transforms.Compose([
        transforms.Resize(256),
        transforms.CenterCrop(224),
        transforms.ToTensor(),
        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
    ]),
}

data_dir = 'data/hymenoptera_data'
image_datasets = {x: datasets.ImageFolder(os.path.join(data_dir, x),
                                          data_transforms[x])
                  for x in ['train', 'val']}
dataloaders = {x: torch.utils.data.DataLoader(image_datasets[x], batch_size=4,
                                             shuffle=True, num_workers=4)
              for x in ['train', 'val']}
dataset_sizes = {x: len(image_datasets[x]) for x in ['train', 'val']}
class_names = image_datasets['train'].classes

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

Visualize a few images

Let’s visualize a few training images so as to understand the data augmentations.

def imshow(inp, title=None):
    """Imshow for Tensor."""
    inp = inp.numpy().transpose((1, 2, 0))
    mean = np.array([0.485, 0.456, 0.406])
    std = np.array([0.229, 0.224, 0.225])
    inp = std * inp + mean
    inp = np.clip(inp, 0, 1)
    plt.imshow(inp)
    if title is not None:
        plt.title(title)
    plt.pause(0.001)  # pause a bit so that plots are updated


# Get a batch of training data
inputs, classes = next(iter(dataloaders['train']))

# Make a grid from batch
out = torchvision.utils.make_grid(inputs)

imshow(out, title=[class_names[x] for x in classes])

Training the model

Now, let’s write a general function to train a model. Here, we will illustrate:

• Scheduling the learning rate
• Saving the best model

In the following, parameter scheduler is an LR scheduler object from torch.optim.lr_scheduler.

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)

        # Each epoch has a training and validation phase
        for phase in ['train', 'val']:
            if phase == 'train':
                scheduler.step()
                model.train()  # Set model to training mode
            else:
                model.eval()   # Set model to evaluate mode

            running_loss = 0.0
            running_corrects = 0

            # Iterate over data.
            for inputs, labels in dataloaders[phase]:
                inputs = inputs.to(device)
                labels = labels.to(device)

                # zero the parameter gradients
                optimizer.zero_grad()

                # forward
                # track history if only in train
                with torch.set_grad_enabled(phase == 'train'):
                    outputs = model(inputs)
                    _, preds = torch.max(outputs, 1)
                    loss = criterion(outputs, labels)

                    # backward + optimize only if in training phase
                    if phase == 'train':
                        loss.backward()
                        optimizer.step()

                # statistics
                running_loss += loss.item() * inputs.size(0)
                running_corrects += torch.sum(preds == labels.data)

            epoch_loss = running_loss / dataset_sizes[phase]
            epoch_acc = running_corrects.double() / dataset_sizes[phase]

            print('{} Loss: {:.4f} Acc: {:.4f}'.format(
                phase, epoch_loss, epoch_acc))

            # deep copy the model
            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))

    # load best model weights
    model.load_state_dict(best_model_wts)
    return model

Visualizing the model predictions

Generic function to display predictions for a few images

def visualize_model(model, num_images=6):
    was_training = model.training
    model.eval()
    images_so_far = 0
    fig = plt.figure()

    with torch.no_grad():
        for i, (inputs, labels) in enumerate(dataloaders['val']):
            inputs = inputs.to(device)
            labels = labels.to(device)

            outputs = model(inputs)
            _, preds = torch.max(outputs, 1)

            for j in range(inputs.size()[0]):
                images_so_far += 1
                ax = plt.subplot(num_images//2, 2, images_so_far)
                ax.axis('off')
                ax.set_title('predicted: {}'.format(class_names[preds[j]]))
                imshow(inputs.cpu().data[j])

                if images_so_far == num_images:
                    model.train(mode=was_training)
                    return
        model.train(mode=was_training)

Finetuning the convnet

Load a pretrained model and reset final fully connected layer.

model_ft = models.resnet18(pretrained=True)
num_ftrs = model_ft.fc.in_features
model_ft.fc = nn.Linear(num_ftrs, 2)

model_ft = model_ft.to(device)

criterion = nn.CrossEntropyLoss()

# Observe that all parameters are being optimized
optimizer_ft = optim.SGD(model_ft.parameters(), lr=0.001, momentum=0.9)

# Decay LR by a factor of 0.1 every 7 epochs
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_ft, step_size=7, gamma=0.1)

Train and evaluate

It should take around 15-25 min on CPU. On GPU though, it takes less than a minute.

model_ft = train_model(model_ft, criterion, optimizer_ft, exp_lr_scheduler,
                       num_epochs=25)

Out:

Epoch 0/24
----------
train Loss: 0.5696 Acc: 0.7295
val Loss: 0.1695 Acc: 0.9412

Epoch 1/24
----------
train Loss: 0.5144 Acc: 0.7951
val Loss: 0.3045 Acc: 0.8889

Epoch 2/24
----------
train Loss: 0.4673 Acc: 0.8156
val Loss: 0.2820 Acc: 0.9020

Epoch 3/24
----------
train Loss: 0.5727 Acc: 0.7746
val Loss: 0.2953 Acc: 0.8627

Epoch 4/24
----------
train Loss: 0.4381 Acc: 0.8115
val Loss: 0.3326 Acc: 0.8562

Epoch 5/24
----------
train Loss: 0.3499 Acc: 0.8525
val Loss: 0.2749 Acc: 0.9150

Epoch 6/24
----------
train Loss: 0.3692 Acc: 0.8730
val Loss: 0.2137 Acc: 0.9150

Epoch 7/24
----------
train Loss: 0.2767 Acc: 0.8975
val Loss: 0.1980 Acc: 0.9281

Epoch 8/24
----------
train Loss: 0.2738 Acc: 0.8893
val Loss: 0.1995 Acc: 0.9216

Epoch 9/24
----------
train Loss: 0.3338 Acc: 0.8689
val Loss: 0.1931 Acc: 0.9346

Epoch 10/24
----------
train Loss: 0.2768 Acc: 0.8811
val Loss: 0.1812 Acc: 0.9346

Epoch 11/24
----------
train Loss: 0.2947 Acc: 0.8443
val Loss: 0.1929 Acc: 0.9412

Epoch 12/24
----------
train Loss: 0.2739 Acc: 0.8730
val Loss: 0.2021 Acc: 0.9150

Epoch 13/24
----------
train Loss: 0.2338 Acc: 0.8975
val Loss: 0.1829 Acc: 0.9412

Epoch 14/24
----------
train Loss: 0.2998 Acc: 0.8648
val Loss: 0.1859 Acc: 0.9346

Epoch 15/24
----------
train Loss: 0.2191 Acc: 0.8934
val Loss: 0.1809 Acc: 0.9346

Epoch 16/24
----------
train Loss: 0.2886 Acc: 0.8770
val Loss: 0.1958 Acc: 0.9150

Epoch 17/24
----------
train Loss: 0.3196 Acc: 0.8525
val Loss: 0.1815 Acc: 0.9412

Epoch 18/24
----------
train Loss: 0.3200 Acc: 0.8566
val Loss: 0.1787 Acc: 0.9412

Epoch 19/24
----------
train Loss: 0.2193 Acc: 0.8975
val Loss: 0.1884 Acc: 0.9346

Epoch 20/24
----------
train Loss: 0.2228 Acc: 0.9221
val Loss: 0.1756 Acc: 0.9412

Epoch 21/24
----------
train Loss: 0.2628 Acc: 0.8852
val Loss: 0.1726 Acc: 0.9412

Epoch 22/24
----------
train Loss: 0.2067 Acc: 0.9180
val Loss: 0.1844 Acc: 0.9346

Epoch 23/24
----------
train Loss: 0.2082 Acc: 0.9344
val Loss: 0.1798 Acc: 0.9412

Epoch 24/24
----------
train Loss: 0.2151 Acc: 0.9139
val Loss: 0.1774 Acc: 0.9346

Training complete in 1m 7s
Best val Acc: 0.941176

visualize_model(model_ft)

ConvNet as fixed feature extractor

Here, we need to freeze all the network except the final layer. We need to set requires_grad == False to freeze the parameters so that the gradients are not computed in backward().

You can read more about this in the documentation here.

model_conv = torchvision.models.resnet18(pretrained=True)
for param in model_conv.parameters():
    param.requires_grad = False

# Parameters of newly constructed modules have requires_grad=True by default
num_ftrs = model_conv.fc.in_features
model_conv.fc = nn.Linear(num_ftrs, 2)

model_conv = model_conv.to(device)

criterion = nn.CrossEntropyLoss()

# Observe that only parameters of final layer are being optimized as
# opposed to before.
optimizer_conv = optim.SGD(model_conv.fc.parameters(), lr=0.001, momentum=0.9)

# Decay LR by a factor of 0.1 every 7 epochs
exp_lr_scheduler = lr_scheduler.StepLR(optimizer_conv, step_size=7, gamma=0.1)

Train and evaluate

On CPU this will take about half the time compared to previous scenario. This is expected as gradients don’t need to be computed for most of the network. However, forward does need to be computed.

model_conv = train_model(model_conv, criterion, optimizer_conv,
                         exp_lr_scheduler, num_epochs=25)

Out:

Epoch 0/24
----------
train Loss: 0.6039 Acc: 0.6516
val Loss: 0.3623 Acc: 0.7974

Epoch 1/24
----------
train Loss: 0.4775 Acc: 0.7582
val Loss: 0.1633 Acc: 0.9477

Epoch 2/24
----------
train Loss: 0.5638 Acc: 0.7787
val Loss: 0.1754 Acc: 0.9346

Epoch 3/24
----------
train Loss: 0.4171 Acc: 0.8361
val Loss: 0.2515 Acc: 0.9216

Epoch 4/24
----------
train Loss: 0.4560 Acc: 0.7828
val Loss: 0.2016 Acc: 0.9281

Epoch 5/24
----------
train Loss: 0.4022 Acc: 0.8197
val Loss: 0.1825 Acc: 0.9477

Epoch 6/24
----------
train Loss: 0.3819 Acc: 0.8320
val Loss: 0.1848 Acc: 0.9412

Epoch 7/24
----------
train Loss: 0.3367 Acc: 0.8443
val Loss: 0.1810 Acc: 0.9542

Epoch 8/24
----------
train Loss: 0.3094 Acc: 0.8648
val Loss: 0.2102 Acc: 0.9477

Epoch 9/24
----------
train Loss: 0.3093 Acc: 0.8566
val Loss: 0.1705 Acc: 0.9542

Epoch 10/24
----------
train Loss: 0.3130 Acc: 0.8484
val Loss: 0.1995 Acc: 0.9412

Epoch 11/24
----------
train Loss: 0.3555 Acc: 0.8361
val Loss: 0.1734 Acc: 0.9542

Epoch 12/24
----------
train Loss: 0.3663 Acc: 0.8238
val Loss: 0.1818 Acc: 0.9542

Epoch 13/24
----------
train Loss: 0.3702 Acc: 0.8320
val Loss: 0.1940 Acc: 0.9542

Epoch 14/24
----------
train Loss: 0.3842 Acc: 0.8361
val Loss: 0.1797 Acc: 0.9542

Epoch 15/24
----------
train Loss: 0.2922 Acc: 0.8648
val Loss: 0.1811 Acc: 0.9542

Epoch 16/24
----------
train Loss: 0.3228 Acc: 0.8566
val Loss: 0.1804 Acc: 0.9477

Epoch 17/24
----------
train Loss: 0.3163 Acc: 0.8525
val Loss: 0.1816 Acc: 0.9542

Epoch 18/24
----------
train Loss: 0.3562 Acc: 0.8238
val Loss: 0.1870 Acc: 0.9542

Epoch 19/24
----------
train Loss: 0.3560 Acc: 0.8361
val Loss: 0.2083 Acc: 0.9346

Epoch 20/24
----------
train Loss: 0.3083 Acc: 0.8893
val Loss: 0.1865 Acc: 0.9542

Epoch 21/24
----------
train Loss: 0.3431 Acc: 0.8566
val Loss: 0.1849 Acc: 0.9542

Epoch 22/24
----------
train Loss: 0.3281 Acc: 0.8566
val Loss: 0.1816 Acc: 0.9542

Epoch 23/24
----------
train Loss: 0.4059 Acc: 0.8115
val Loss: 0.1651 Acc: 0.9542

Epoch 24/24
----------
train Loss: 0.2590 Acc: 0.8730
val Loss: 0.2005 Acc: 0.9412

Training complete in 0m 35s
Best val Acc: 0.954248

visualize_model(model_conv)

plt.ioff()
plt.show()