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 0x7f5d19d40e20>

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]
 65%|######4   | 29.0M/44.7M [00:00<00:00, 303MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 290MB/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: 0.5880 Acc: 0.6844
val Loss: 0.2693 Acc: 0.9085

Epoch 1/24
----------
train Loss: 0.4562 Acc: 0.8115
val Loss: 0.3616 Acc: 0.8562

Epoch 2/24
----------
train Loss: 0.6907 Acc: 0.7090
val Loss: 0.4877 Acc: 0.8497

Epoch 3/24
----------
train Loss: 0.3737 Acc: 0.8320
val Loss: 0.4975 Acc: 0.7974

Epoch 4/24
----------
train Loss: 0.4644 Acc: 0.7992
val Loss: 0.4147 Acc: 0.8431

Epoch 5/24
----------
train Loss: 0.4558 Acc: 0.8033
val Loss: 0.2843 Acc: 0.8954

Epoch 6/24
----------
train Loss: 0.4574 Acc: 0.8197
val Loss: 0.7101 Acc: 0.8039

Epoch 7/24
----------
train Loss: 0.4203 Acc: 0.8402
val Loss: 0.3246 Acc: 0.8758

Epoch 8/24
----------
train Loss: 0.3073 Acc: 0.8607
val Loss: 0.3257 Acc: 0.8758

Epoch 9/24
----------
train Loss: 0.2925 Acc: 0.8689
val Loss: 0.2790 Acc: 0.9020

Epoch 10/24
----------
train Loss: 0.3599 Acc: 0.8156
val Loss: 0.2768 Acc: 0.8954

Epoch 11/24
----------
train Loss: 0.3157 Acc: 0.8770
val Loss: 0.2334 Acc: 0.9150

Epoch 12/24
----------
train Loss: 0.2942 Acc: 0.8361
val Loss: 0.2527 Acc: 0.9085

Epoch 13/24
----------
train Loss: 0.3238 Acc: 0.8566
val Loss: 0.2555 Acc: 0.8954

Epoch 14/24
----------
train Loss: 0.2730 Acc: 0.8730
val Loss: 0.2462 Acc: 0.8889

Epoch 15/24
----------
train Loss: 0.2862 Acc: 0.8852
val Loss: 0.3031 Acc: 0.8693

Epoch 16/24
----------
train Loss: 0.2604 Acc: 0.8893
val Loss: 0.2469 Acc: 0.9085

Epoch 17/24
----------
train Loss: 0.3466 Acc: 0.8361
val Loss: 0.2247 Acc: 0.9216

Epoch 18/24
----------
train Loss: 0.2829 Acc: 0.8852
val Loss: 0.2195 Acc: 0.9216

Epoch 19/24
----------
train Loss: 0.3518 Acc: 0.8689
val Loss: 0.2323 Acc: 0.9281

Epoch 20/24
----------
train Loss: 0.2589 Acc: 0.8852
val Loss: 0.2254 Acc: 0.9216

Epoch 21/24
----------
train Loss: 0.3505 Acc: 0.8197
val Loss: 0.2234 Acc: 0.9281

Epoch 22/24
----------
train Loss: 0.2537 Acc: 0.8975
val Loss: 0.2361 Acc: 0.9216

Epoch 23/24
----------
train Loss: 0.2935 Acc: 0.8689
val Loss: 0.3165 Acc: 0.8693

Epoch 24/24
----------
train Loss: 0.2978 Acc: 0.8689
val Loss: 0.2671 Acc: 0.8824

Training complete in 0m 35s
Best val Acc: 0.928105

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: 0.6808 Acc: 0.6475
val Loss: 0.2835 Acc: 0.8889

Epoch 1/24
----------
train Loss: 0.4422 Acc: 0.7869
val Loss: 0.3210 Acc: 0.8497

Epoch 2/24
----------
train Loss: 0.4822 Acc: 0.7705
val Loss: 0.2221 Acc: 0.9150

Epoch 3/24
----------
train Loss: 0.5183 Acc: 0.7910
val Loss: 0.1868 Acc: 0.9477

Epoch 4/24
----------
train Loss: 0.4816 Acc: 0.7951
val Loss: 0.4276 Acc: 0.8431

Epoch 5/24
----------
train Loss: 0.4767 Acc: 0.8115
val Loss: 0.1970 Acc: 0.9281

Epoch 6/24
----------
train Loss: 0.3801 Acc: 0.8279
val Loss: 0.2243 Acc: 0.9150

Epoch 7/24
----------
train Loss: 0.3645 Acc: 0.8484
val Loss: 0.1903 Acc: 0.9477

Epoch 8/24
----------
train Loss: 0.4290 Acc: 0.8115
val Loss: 0.2051 Acc: 0.9412

Epoch 9/24
----------
train Loss: 0.3243 Acc: 0.8566
val Loss: 0.2123 Acc: 0.9281

Epoch 10/24
----------
train Loss: 0.3366 Acc: 0.8361
val Loss: 0.1902 Acc: 0.9477

Epoch 11/24
----------
train Loss: 0.3678 Acc: 0.8238
val Loss: 0.2161 Acc: 0.9346

Epoch 12/24
----------
train Loss: 0.3707 Acc: 0.8443
val Loss: 0.1738 Acc: 0.9477

Epoch 13/24
----------
train Loss: 0.2666 Acc: 0.9098
val Loss: 0.2084 Acc: 0.9346

Epoch 14/24
----------
train Loss: 0.3166 Acc: 0.8607
val Loss: 0.2143 Acc: 0.9477

Epoch 15/24
----------
train Loss: 0.3091 Acc: 0.8648
val Loss: 0.2358 Acc: 0.9150

Epoch 16/24
----------
train Loss: 0.3598 Acc: 0.8361
val Loss: 0.1841 Acc: 0.9412

Epoch 17/24
----------
train Loss: 0.3442 Acc: 0.8443
val Loss: 0.2135 Acc: 0.9281

Epoch 18/24
----------
train Loss: 0.2933 Acc: 0.8689
val Loss: 0.2365 Acc: 0.9085

Epoch 19/24
----------
train Loss: 0.3038 Acc: 0.8730
val Loss: 0.1898 Acc: 0.9477

Epoch 20/24
----------
train Loss: 0.3680 Acc: 0.8402
val Loss: 0.2048 Acc: 0.9346

Epoch 21/24
----------
train Loss: 0.2950 Acc: 0.8770
val Loss: 0.2137 Acc: 0.9281

Epoch 22/24
----------
train Loss: 0.2818 Acc: 0.8607
val Loss: 0.1791 Acc: 0.9477

Epoch 23/24
----------
train Loss: 0.3047 Acc: 0.8443
val Loss: 0.2101 Acc: 0.9346

Epoch 24/24
----------
train Loss: 0.3732 Acc: 0.8197
val Loss: 0.2032 Acc: 0.9281

Training complete in 0m 28s
Best val Acc: 0.947712

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.