Transfer Learning for Computer Vision Tutorial

Author: Sasank Chilamkurthy

In this tutorial, you will learn how to train a convolutional neural network for image classification 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':
                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)
            if phase == 'train':
                scheduler.step()

            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
# Here the size of each output sample is set to 2.
# Alternatively, it can be generalized to nn.Linear(num_ftrs, len(class_names)).
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: 0.6702 Acc: 0.7090
val Loss: 0.2730 Acc: 0.9020

Epoch 1/24
----------
train Loss: 0.6729 Acc: 0.7541
val Loss: 0.3299 Acc: 0.8954

Epoch 2/24
----------
train Loss: 0.6817 Acc: 0.7336
val Loss: 0.3307 Acc: 0.8497

Epoch 3/24
----------
train Loss: 0.4983 Acc: 0.7787
val Loss: 0.1728 Acc: 0.9346

Epoch 4/24
----------
train Loss: 0.5486 Acc: 0.7541
val Loss: 0.3014 Acc: 0.8301

Epoch 5/24
----------
train Loss: 0.4349 Acc: 0.7992
val Loss: 0.4873 Acc: 0.8170

Epoch 6/24
----------
train Loss: 0.4801 Acc: 0.8156
val Loss: 0.2955 Acc: 0.8889

Epoch 7/24
----------
train Loss: 0.3034 Acc: 0.8770
val Loss: 0.2256 Acc: 0.9216

Epoch 8/24
----------
train Loss: 0.3516 Acc: 0.8689
val Loss: 0.2475 Acc: 0.9150

Epoch 9/24
----------
train Loss: 0.3342 Acc: 0.8566
val Loss: 0.2038 Acc: 0.9346

Epoch 10/24
----------
train Loss: 0.2998 Acc: 0.8648
val Loss: 0.2882 Acc: 0.9020

Epoch 11/24
----------
train Loss: 0.2981 Acc: 0.8852
val Loss: 0.1921 Acc: 0.9346

Epoch 12/24
----------
train Loss: 0.3717 Acc: 0.8197
val Loss: 0.2140 Acc: 0.9216

Epoch 13/24
----------
train Loss: 0.4508 Acc: 0.8033
val Loss: 0.2785 Acc: 0.9020

Epoch 14/24
----------
train Loss: 0.3103 Acc: 0.8770
val Loss: 0.2123 Acc: 0.9216

Epoch 15/24
----------
train Loss: 0.2257 Acc: 0.9057
val Loss: 0.2388 Acc: 0.9150

Epoch 16/24
----------
train Loss: 0.2491 Acc: 0.8811
val Loss: 0.2254 Acc: 0.9150

Epoch 17/24
----------
train Loss: 0.2895 Acc: 0.8689
val Loss: 0.2118 Acc: 0.9346

Epoch 18/24
----------
train Loss: 0.2746 Acc: 0.8852
val Loss: 0.2142 Acc: 0.9281

Epoch 19/24
----------
train Loss: 0.2608 Acc: 0.8852
val Loss: 0.2696 Acc: 0.8889

Epoch 20/24
----------
train Loss: 0.3047 Acc: 0.8811
val Loss: 0.1961 Acc: 0.9216

Epoch 21/24
----------
train Loss: 0.2937 Acc: 0.8689
val Loss: 0.2078 Acc: 0.9412

Epoch 22/24
----------
train Loss: 0.2201 Acc: 0.9139
val Loss: 0.2073 Acc: 0.9346

Epoch 23/24
----------
train Loss: 0.2855 Acc: 0.8852
val Loss: 0.2200 Acc: 0.9281

Epoch 24/24
----------
train Loss: 0.2792 Acc: 0.8770
val Loss: 0.2122 Acc: 0.9281

Training complete in 1m 10s
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: 0.5994 Acc: 0.7131
val Loss: 0.2415 Acc: 0.9281

Epoch 1/24
----------
train Loss: 0.6533 Acc: 0.6885
val Loss: 0.4124 Acc: 0.7778

Epoch 2/24
----------
train Loss: 0.4680 Acc: 0.7992
val Loss: 0.2008 Acc: 0.9281

Epoch 3/24
----------
train Loss: 0.4567 Acc: 0.7541
val Loss: 0.1943 Acc: 0.9477

Epoch 4/24
----------
train Loss: 0.6596 Acc: 0.7336
val Loss: 0.3974 Acc: 0.8693

Epoch 5/24
----------
train Loss: 0.5504 Acc: 0.7828
val Loss: 0.1689 Acc: 0.9477

Epoch 6/24
----------
train Loss: 0.4323 Acc: 0.8402
val Loss: 0.1648 Acc: 0.9477

Epoch 7/24
----------
train Loss: 0.4636 Acc: 0.8033
val Loss: 0.1968 Acc: 0.9346

Epoch 8/24
----------
train Loss: 0.3015 Acc: 0.8566
val Loss: 0.1650 Acc: 0.9412

Epoch 9/24
----------
train Loss: 0.3832 Acc: 0.8238
val Loss: 0.1748 Acc: 0.9477

Epoch 10/24
----------
train Loss: 0.3466 Acc: 0.8443
val Loss: 0.1738 Acc: 0.9477

Epoch 11/24
----------
train Loss: 0.3891 Acc: 0.8074
val Loss: 0.1563 Acc: 0.9477

Epoch 12/24
----------
train Loss: 0.3649 Acc: 0.8279
val Loss: 0.1812 Acc: 0.9477

Epoch 13/24
----------
train Loss: 0.3372 Acc: 0.8607
val Loss: 0.1614 Acc: 0.9542

Epoch 14/24
----------
train Loss: 0.3300 Acc: 0.8689
val Loss: 0.2195 Acc: 0.9150

Epoch 15/24
----------
train Loss: 0.3547 Acc: 0.8402
val Loss: 0.1863 Acc: 0.9412

Epoch 16/24
----------
train Loss: 0.3495 Acc: 0.8402
val Loss: 0.1706 Acc: 0.9542

Epoch 17/24
----------
train Loss: 0.3112 Acc: 0.8443
val Loss: 0.1676 Acc: 0.9477

Epoch 18/24
----------
train Loss: 0.3550 Acc: 0.8443
val Loss: 0.1642 Acc: 0.9477

Epoch 19/24
----------
train Loss: 0.3773 Acc: 0.8279
val Loss: 0.1751 Acc: 0.9477

Epoch 20/24
----------
train Loss: 0.3684 Acc: 0.8402
val Loss: 0.1606 Acc: 0.9477

Epoch 21/24
----------
train Loss: 0.3514 Acc: 0.8607
val Loss: 0.2011 Acc: 0.9346

Epoch 22/24
----------
train Loss: 0.3707 Acc: 0.8320
val Loss: 0.1684 Acc: 0.9477

Epoch 23/24
----------
train Loss: 0.3505 Acc: 0.8197
val Loss: 0.1758 Acc: 0.9412

Epoch 24/24
----------
train Loss: 0.3540 Acc: 0.8361
val Loss: 0.1667 Acc: 0.9477

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

visualize_model(model_conv)

plt.ioff()
plt.show()

Further Learning

If you would like to learn more about the applications of transfer learning, checkout our Quantized Transfer Learning for Computer Vision Tutorial.