beginner/transfer_learning_tutorial

Run in Google Colab

Colab

Download Notebook

Notebook

View on GitHub

GitHub

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 0x7f0a0e836740>

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]
 88%|████████▊ | 39.2M/44.7M [00:00<00:00, 411MB/s]
100%|██████████| 44.7M/44.7M [00:00<00:00, 414MB/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: 0.6010 Acc: 0.6844
val Loss: 0.2630 Acc: 0.9020

Epoch 1/24
----------
train Loss: 0.5260 Acc: 0.7459
val Loss: 0.1744 Acc: 0.9150

Epoch 2/24
----------
train Loss: 0.6230 Acc: 0.7664
val Loss: 0.2028 Acc: 0.9085

Epoch 3/24
----------
train Loss: 0.5234 Acc: 0.7787
val Loss: 0.2164 Acc: 0.9281

Epoch 4/24
----------
train Loss: 0.6832 Acc: 0.7418
val Loss: 0.2535 Acc: 0.8954

Epoch 5/24
----------
train Loss: 0.4053 Acc: 0.8320
val Loss: 0.2064 Acc: 0.9281

Epoch 6/24
----------
train Loss: 0.3929 Acc: 0.8238
val Loss: 0.4418 Acc: 0.8497

Epoch 7/24
----------
train Loss: 0.4170 Acc: 0.8074
val Loss: 0.2969 Acc: 0.8824

Epoch 8/24
----------
train Loss: 0.3056 Acc: 0.8689
val Loss: 0.2911 Acc: 0.8824

Epoch 9/24
----------
train Loss: 0.2768 Acc: 0.9016
val Loss: 0.2468 Acc: 0.9085

Epoch 10/24
----------
train Loss: 0.3056 Acc: 0.8566
val Loss: 0.2464 Acc: 0.9020

Epoch 11/24
----------
train Loss: 0.2533 Acc: 0.8893
val Loss: 0.2471 Acc: 0.9085

Epoch 12/24
----------
train Loss: 0.3200 Acc: 0.8730
val Loss: 0.2437 Acc: 0.9085

Epoch 13/24
----------
train Loss: 0.2498 Acc: 0.8893
val Loss: 0.2196 Acc: 0.9150

Epoch 14/24
----------
train Loss: 0.2961 Acc: 0.8648
val Loss: 0.2224 Acc: 0.9216

Epoch 15/24
----------
train Loss: 0.2773 Acc: 0.8811
val Loss: 0.2279 Acc: 0.9216

Epoch 16/24
----------
train Loss: 0.4074 Acc: 0.8279
val Loss: 0.2435 Acc: 0.9020

Epoch 17/24
----------
train Loss: 0.2832 Acc: 0.8975
val Loss: 0.2208 Acc: 0.9150

Epoch 18/24
----------
train Loss: 0.2351 Acc: 0.9139
val Loss: 0.2395 Acc: 0.9150

Epoch 19/24
----------
train Loss: 0.3115 Acc: 0.8566
val Loss: 0.2235 Acc: 0.9150

Epoch 20/24
----------
train Loss: 0.2861 Acc: 0.8770
val Loss: 0.2343 Acc: 0.9085

Epoch 21/24
----------
train Loss: 0.2282 Acc: 0.9057
val Loss: 0.2307 Acc: 0.9281

Epoch 22/24
----------
train Loss: 0.2379 Acc: 0.9016
val Loss: 0.2198 Acc: 0.9346

Epoch 23/24
----------
train Loss: 0.2652 Acc: 0.9098
val Loss: 0.2362 Acc: 0.9085

Epoch 24/24
----------
train Loss: 0.2274 Acc: 0.9098
val Loss: 0.2364 Acc: 0.9150

Training complete in 0m 34s
Best val Acc: 0.934641

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: 0.6500 Acc: 0.6311
val Loss: 0.1972 Acc: 0.9608

Epoch 1/24
----------
train Loss: 0.4726 Acc: 0.7828
val Loss: 0.1576 Acc: 0.9542

Epoch 2/24
----------
train Loss: 0.5731 Acc: 0.7418
val Loss: 0.1581 Acc: 0.9542

Epoch 3/24
----------
train Loss: 0.3397 Acc: 0.8484
val Loss: 0.2111 Acc: 0.9346

Epoch 4/24
----------
train Loss: 0.6174 Acc: 0.7828
val Loss: 0.1893 Acc: 0.9412

Epoch 5/24
----------
train Loss: 0.5276 Acc: 0.7910
val Loss: 0.2630 Acc: 0.9150

Epoch 6/24
----------
train Loss: 0.5067 Acc: 0.8033
val Loss: 0.2675 Acc: 0.9020

Epoch 7/24
----------
train Loss: 0.3642 Acc: 0.8443
val Loss: 0.1356 Acc: 0.9673

Epoch 8/24
----------
train Loss: 0.2843 Acc: 0.8770
val Loss: 0.1571 Acc: 0.9608

Epoch 9/24
----------
train Loss: 0.3888 Acc: 0.8443
val Loss: 0.1969 Acc: 0.9281

Epoch 10/24
----------
train Loss: 0.3848 Acc: 0.8197
val Loss: 0.1855 Acc: 0.9608

Epoch 11/24
----------
train Loss: 0.3933 Acc: 0.8361
val Loss: 0.1836 Acc: 0.9477

Epoch 12/24
----------
train Loss: 0.4084 Acc: 0.8320
val Loss: 0.1899 Acc: 0.9542

Epoch 13/24
----------
train Loss: 0.3889 Acc: 0.8238
val Loss: 0.1745 Acc: 0.9542

Epoch 14/24
----------
train Loss: 0.2673 Acc: 0.8934
val Loss: 0.1802 Acc: 0.9477

Epoch 15/24
----------
train Loss: 0.2769 Acc: 0.8811
val Loss: 0.1832 Acc: 0.9542

Epoch 16/24
----------
train Loss: 0.3020 Acc: 0.8770
val Loss: 0.1727 Acc: 0.9542

Epoch 17/24
----------
train Loss: 0.3457 Acc: 0.8566
val Loss: 0.1721 Acc: 0.9608

Epoch 18/24
----------
train Loss: 0.3702 Acc: 0.8566
val Loss: 0.1546 Acc: 0.9608

Epoch 19/24
----------
train Loss: 0.5122 Acc: 0.7828
val Loss: 0.1814 Acc: 0.9542

Epoch 20/24
----------
train Loss: 0.2892 Acc: 0.8770
val Loss: 0.1712 Acc: 0.9608

Epoch 21/24
----------
train Loss: 0.3547 Acc: 0.8525
val Loss: 0.1724 Acc: 0.9608

Epoch 22/24
----------
train Loss: 0.3297 Acc: 0.8811
val Loss: 0.1769 Acc: 0.9542

Epoch 23/24
----------
train Loss: 0.3066 Acc: 0.8730
val Loss: 0.1856 Acc: 0.9542

Epoch 24/24
----------
train Loss: 0.3817 Acc: 0.8402
val Loss: 0.1596 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.