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 0x7fb65bba4e20>

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]
 89%|########9 | 39.9M/44.7M [00:00<00:00, 417MB/s]
100%|##########| 44.7M/44.7M [00:00<00:00, 418MB/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: 0.6274 Acc: 0.6885
val Loss: 0.2350 Acc: 0.9085

Epoch 1/24
----------
train Loss: 0.4421 Acc: 0.8074
val Loss: 0.3403 Acc: 0.8758

Epoch 2/24
----------
train Loss: 0.4541 Acc: 0.8074
val Loss: 0.2733 Acc: 0.8954

Epoch 3/24
----------
train Loss: 0.4344 Acc: 0.7992
val Loss: 0.5466 Acc: 0.7778

Epoch 4/24
----------
train Loss: 0.7619 Acc: 0.7377
val Loss: 0.5163 Acc: 0.7908

Epoch 5/24
----------
train Loss: 0.4621 Acc: 0.7910
val Loss: 0.4879 Acc: 0.7908

Epoch 6/24
----------
train Loss: 0.4671 Acc: 0.8279
val Loss: 0.5458 Acc: 0.8366

Epoch 7/24
----------
train Loss: 0.2781 Acc: 0.8566
val Loss: 0.2782 Acc: 0.8889

Epoch 8/24
----------
train Loss: 0.2935 Acc: 0.8730
val Loss: 0.2785 Acc: 0.9020

Epoch 9/24
----------
train Loss: 0.3242 Acc: 0.8484
val Loss: 0.2610 Acc: 0.9085

Epoch 10/24
----------
train Loss: 0.3817 Acc: 0.8361
val Loss: 0.3016 Acc: 0.8954

Epoch 11/24
----------
train Loss: 0.3288 Acc: 0.8484
val Loss: 0.2902 Acc: 0.8954

Epoch 12/24
----------
train Loss: 0.3229 Acc: 0.8730
val Loss: 0.3495 Acc: 0.8562

Epoch 13/24
----------
train Loss: 0.2757 Acc: 0.8730
val Loss: 0.2860 Acc: 0.9216

Epoch 14/24
----------
train Loss: 0.2222 Acc: 0.9303
val Loss: 0.2880 Acc: 0.8954

Epoch 15/24
----------
train Loss: 0.3097 Acc: 0.8689
val Loss: 0.2694 Acc: 0.9085

Epoch 16/24
----------
train Loss: 0.3802 Acc: 0.8320
val Loss: 0.3044 Acc: 0.8889

Epoch 17/24
----------
train Loss: 0.3431 Acc: 0.8566
val Loss: 0.2824 Acc: 0.9150

Epoch 18/24
----------
train Loss: 0.3043 Acc: 0.8607
val Loss: 0.3062 Acc: 0.9020

Epoch 19/24
----------
train Loss: 0.2844 Acc: 0.8770
val Loss: 0.2803 Acc: 0.9085

Epoch 20/24
----------
train Loss: 0.3610 Acc: 0.8320
val Loss: 0.3398 Acc: 0.8693

Epoch 21/24
----------
train Loss: 0.2555 Acc: 0.8811
val Loss: 0.2953 Acc: 0.9085

Epoch 22/24
----------
train Loss: 0.3109 Acc: 0.8607
val Loss: 0.2993 Acc: 0.9020

Epoch 23/24
----------
train Loss: 0.2399 Acc: 0.8975
val Loss: 0.3416 Acc: 0.8627

Epoch 24/24
----------
train Loss: 0.3136 Acc: 0.8648
val Loss: 0.3023 Acc: 0.9020

Training complete in 0m 35s
Best val Acc: 0.921569

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: 0.6947 Acc: 0.6475
val Loss: 0.2573 Acc: 0.9020

Epoch 1/24
----------
train Loss: 0.5811 Acc: 0.7623
val Loss: 0.1844 Acc: 0.9542

Epoch 2/24
----------
train Loss: 0.4425 Acc: 0.7910
val Loss: 0.1673 Acc: 0.9477

Epoch 3/24
----------
train Loss: 0.5132 Acc: 0.7910
val Loss: 0.2252 Acc: 0.9216

Epoch 4/24
----------
train Loss: 0.7139 Acc: 0.7418
val Loss: 0.2706 Acc: 0.9085

Epoch 5/24
----------
train Loss: 0.3668 Acc: 0.8361
val Loss: 0.1949 Acc: 0.9542

Epoch 6/24
----------
train Loss: 0.4035 Acc: 0.8484
val Loss: 0.2148 Acc: 0.9346

Epoch 7/24
----------
train Loss: 0.4813 Acc: 0.7828
val Loss: 0.2735 Acc: 0.8954

Epoch 8/24
----------
train Loss: 0.3637 Acc: 0.8525
val Loss: 0.1963 Acc: 0.9412

Epoch 9/24
----------
train Loss: 0.3372 Acc: 0.8607
val Loss: 0.1897 Acc: 0.9477

Epoch 10/24
----------
train Loss: 0.3948 Acc: 0.8156
val Loss: 0.1991 Acc: 0.9477

Epoch 11/24
----------
train Loss: 0.4904 Acc: 0.7705
val Loss: 0.1946 Acc: 0.9477

Epoch 12/24
----------
train Loss: 0.3374 Acc: 0.8443
val Loss: 0.2012 Acc: 0.9346

Epoch 13/24
----------
train Loss: 0.3086 Acc: 0.8730
val Loss: 0.2220 Acc: 0.9150

Epoch 14/24
----------
train Loss: 0.3638 Acc: 0.8443
val Loss: 0.2323 Acc: 0.9216

Epoch 15/24
----------
train Loss: 0.4637 Acc: 0.7910
val Loss: 0.1948 Acc: 0.9412

Epoch 16/24
----------
train Loss: 0.2940 Acc: 0.8689
val Loss: 0.2002 Acc: 0.9412

Epoch 17/24
----------
train Loss: 0.3700 Acc: 0.8279
val Loss: 0.2162 Acc: 0.9216

Epoch 18/24
----------
train Loss: 0.3036 Acc: 0.8607
val Loss: 0.2247 Acc: 0.9216

Epoch 19/24
----------
train Loss: 0.3513 Acc: 0.8443
val Loss: 0.1918 Acc: 0.9542

Epoch 20/24
----------
train Loss: 0.3348 Acc: 0.8648
val Loss: 0.2142 Acc: 0.9281

Epoch 21/24
----------
train Loss: 0.3701 Acc: 0.8402
val Loss: 0.2064 Acc: 0.9412

Epoch 22/24
----------
train Loss: 0.4008 Acc: 0.7992
val Loss: 0.1888 Acc: 0.9542

Epoch 23/24
----------
train Loss: 0.3190 Acc: 0.8689
val Loss: 0.1857 Acc: 0.9542

Epoch 24/24
----------
train Loss: 0.3617 Acc: 0.8402
val Loss: 0.2369 Acc: 0.9150

Training complete in 0m 27s
Best val Acc: 0.954248

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.