Transfer Learning for Computer Vision Tutorial

Created On: Mar 24, 2017 | Last Updated: Jan 27, 2025 | Last Verified: Nov 05, 2024

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 initialization, 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

import torch
import torch.nn as nn
import torch.optim as optim
from torch.optim import lr_scheduler
import torch.backends.cudnn as cudnn
import numpy as np
import torchvision
from torchvision import datasets, models, transforms
import matplotlib.pyplot as plt
import time
import os
from PIL import Image
from tempfile import TemporaryDirectory

cudnn.benchmark = True
plt.ion()   # interactive mode

<contextlib.ExitStack object at 0x7f1aea588b50>

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

# We want to be able to train our model on an `accelerator <https://pytorch.org/docs/stable/torch.html#accelerators>`__
# such as CUDA, MPS, MTIA, or XPU. If the current accelerator is available, we will use it. Otherwise, we use the CPU.

device = torch.accelerator.current_accelerator().type if torch.accelerator.is_available() else "cpu"
print(f"Using {device} device")

Using cuda device

Visualize a few images

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

def imshow(inp, title=None):
    """Display image 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()

    # Create a temporary directory to save training checkpoints
    with TemporaryDirectory() as tempdir:
        best_model_params_path = os.path.join(tempdir, 'best_model_params.pt')

        torch.save(model.state_dict(), best_model_params_path)
        best_acc = 0.0

        for epoch in range(num_epochs):
            print(f'Epoch {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(f'{phase} Loss: {epoch_loss:.4f} Acc: {epoch_acc:.4f}')

                # deep copy the model
                if phase == 'val' and epoch_acc > best_acc:
                    best_acc = epoch_acc
                    torch.save(model.state_dict(), best_model_params_path)

            print()

        time_elapsed = time.time() - since
        print(f'Training complete in {time_elapsed // 60:.0f}m {time_elapsed % 60:.0f}s')
        print(f'Best val Acc: {best_acc:4f}')

        # load best model weights
        model.load_state_dict(torch.load(best_model_params_path, weights_only=True))
    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(f'predicted: {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(weights='IMAGENET1K_V1')
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)

Downloading: "https://download.pytorch.org/models/resnet18-f37072fd.pth" to /var/lib/ci-user/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth

  0%|          | 0.00/44.7M [00:00<?, ?B/s]
 83%|########3 | 37.2M/44.7M [00:00<00:00, 388MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 366MB/s]

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)

Epoch 0/24
----------
train Loss: 0.4927 Acc: 0.7418
val Loss: 0.2132 Acc: 0.9216

Epoch 1/24
----------
train Loss: 0.6731 Acc: 0.7418
val Loss: 0.3122 Acc: 0.9020

Epoch 2/24
----------
train Loss: 0.5280 Acc: 0.7910
val Loss: 0.3259 Acc: 0.8758

Epoch 3/24
----------
train Loss: 0.5601 Acc: 0.7664
val Loss: 0.7301 Acc: 0.7386

Epoch 4/24
----------
train Loss: 0.3613 Acc: 0.8402
val Loss: 0.2888 Acc: 0.9281

Epoch 5/24
----------
train Loss: 0.4747 Acc: 0.8279
val Loss: 0.2762 Acc: 0.9150

Epoch 6/24
----------
train Loss: 0.3398 Acc: 0.8279
val Loss: 0.2883 Acc: 0.8954

Epoch 7/24
----------
train Loss: 0.3068 Acc: 0.8852
val Loss: 0.2627 Acc: 0.9150

Epoch 8/24
----------
train Loss: 0.2685 Acc: 0.8811
val Loss: 0.2286 Acc: 0.9150

Epoch 9/24
----------
train Loss: 0.2895 Acc: 0.8730
val Loss: 0.2155 Acc: 0.9281

Epoch 10/24
----------
train Loss: 0.3169 Acc: 0.8402
val Loss: 0.2007 Acc: 0.9281

Epoch 11/24
----------
train Loss: 0.2925 Acc: 0.8566
val Loss: 0.2033 Acc: 0.9346

Epoch 12/24
----------
train Loss: 0.2820 Acc: 0.8730
val Loss: 0.2837 Acc: 0.9150

Epoch 13/24
----------
train Loss: 0.2501 Acc: 0.8934
val Loss: 0.2292 Acc: 0.9216

Epoch 14/24
----------
train Loss: 0.2661 Acc: 0.8934
val Loss: 0.2078 Acc: 0.9346

Epoch 15/24
----------
train Loss: 0.1473 Acc: 0.9467
val Loss: 0.2254 Acc: 0.9216

Epoch 16/24
----------
train Loss: 0.3126 Acc: 0.8566
val Loss: 0.2051 Acc: 0.9477

Epoch 17/24
----------
train Loss: 0.3284 Acc: 0.8730
val Loss: 0.2193 Acc: 0.9216

Epoch 18/24
----------
train Loss: 0.3353 Acc: 0.8648
val Loss: 0.2099 Acc: 0.9346

Epoch 19/24
----------
train Loss: 0.2796 Acc: 0.8893
val Loss: 0.2069 Acc: 0.9216

Epoch 20/24
----------
train Loss: 0.3135 Acc: 0.8484
val Loss: 0.1822 Acc: 0.9346

Epoch 21/24
----------
train Loss: 0.2407 Acc: 0.9016
val Loss: 0.1965 Acc: 0.9412

Epoch 22/24
----------
train Loss: 0.2321 Acc: 0.9057
val Loss: 0.2072 Acc: 0.9281

Epoch 23/24
----------
train Loss: 0.2326 Acc: 0.9221
val Loss: 0.2246 Acc: 0.9216

Epoch 24/24
----------
train Loss: 0.1989 Acc: 0.9262
val Loss: 0.2025 Acc: 0.9346

Training complete in 0m 34s
Best val Acc: 0.947712

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(weights='IMAGENET1K_V1')
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)

Epoch 0/24
----------
train Loss: 0.5950 Acc: 0.6680
val Loss: 0.3172 Acc: 0.8627

Epoch 1/24
----------
train Loss: 0.4461 Acc: 0.7910
val Loss: 0.3242 Acc: 0.8627

Epoch 2/24
----------
train Loss: 0.4553 Acc: 0.7951
val Loss: 0.1822 Acc: 0.9412

Epoch 3/24
----------
train Loss: 0.4941 Acc: 0.7582
val Loss: 0.4157 Acc: 0.8301

Epoch 4/24
----------
train Loss: 0.6646 Acc: 0.7336
val Loss: 0.1680 Acc: 0.9542

Epoch 5/24
----------
train Loss: 0.2988 Acc: 0.8648
val Loss: 0.1598 Acc: 0.9542

Epoch 6/24
----------
train Loss: 0.7342 Acc: 0.7090
val Loss: 0.4875 Acc: 0.8301

Epoch 7/24
----------
train Loss: 0.5252 Acc: 0.8033
val Loss: 0.1692 Acc: 0.9673

Epoch 8/24
----------
train Loss: 0.4300 Acc: 0.7910
val Loss: 0.1685 Acc: 0.9542

Epoch 9/24
----------
train Loss: 0.3796 Acc: 0.8156
val Loss: 0.1706 Acc: 0.9673

Epoch 10/24
----------
train Loss: 0.3611 Acc: 0.8320
val Loss: 0.1811 Acc: 0.9412

Epoch 11/24
----------
train Loss: 0.3332 Acc: 0.8525
val Loss: 0.1727 Acc: 0.9542

Epoch 12/24
----------
train Loss: 0.3334 Acc: 0.8566
val Loss: 0.1724 Acc: 0.9608

Epoch 13/24
----------
train Loss: 0.3273 Acc: 0.8361
val Loss: 0.1700 Acc: 0.9608

Epoch 14/24
----------
train Loss: 0.2639 Acc: 0.9016
val Loss: 0.2115 Acc: 0.9477

Epoch 15/24
----------
train Loss: 0.3552 Acc: 0.8443
val Loss: 0.1600 Acc: 0.9542

Epoch 16/24
----------
train Loss: 0.2710 Acc: 0.9057
val Loss: 0.1647 Acc: 0.9542

Epoch 17/24
----------
train Loss: 0.3824 Acc: 0.8279
val Loss: 0.1717 Acc: 0.9542

Epoch 18/24
----------
train Loss: 0.3604 Acc: 0.8361
val Loss: 0.1686 Acc: 0.9542

Epoch 19/24
----------
train Loss: 0.2983 Acc: 0.8648
val Loss: 0.1682 Acc: 0.9542

Epoch 20/24
----------
train Loss: 0.2557 Acc: 0.8934
val Loss: 0.2204 Acc: 0.9281

Epoch 21/24
----------
train Loss: 0.3688 Acc: 0.8648
val Loss: 0.1784 Acc: 0.9412

Epoch 22/24
----------
train Loss: 0.3745 Acc: 0.8525
val Loss: 0.2285 Acc: 0.9150

Epoch 23/24
----------
train Loss: 0.3280 Acc: 0.8566
val Loss: 0.1806 Acc: 0.9477

Epoch 24/24
----------
train Loss: 0.4019 Acc: 0.8402
val Loss: 0.1697 Acc: 0.9608

Training complete in 0m 27s
Best val Acc: 0.967320

visualize_model(model_conv)

plt.ioff()
plt.show()

Inference on custom images

Use the trained model to make predictions on custom images and visualize the predicted class labels along with the images.

def visualize_model_predictions(model,img_path):
    was_training = model.training
    model.eval()

    img = Image.open(img_path)
    img = data_transforms['val'](img)
    img = img.unsqueeze(0)
    img = img.to(device)

    with torch.no_grad():
        outputs = model(img)
        _, preds = torch.max(outputs, 1)

        ax = plt.subplot(2,2,1)
        ax.axis('off')
        ax.set_title(f'Predicted: {class_names[preds[0]]}')
        imshow(img.cpu().data[0])

        model.train(mode=was_training)

visualize_model_predictions(
    model_conv,
    img_path='data/hymenoptera_data/val/bees/72100438_73de9f17af.jpg'
)

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.