Note
Go to the end to download the full example code.
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 0x7f084e4b2800>
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])
![['ants', 'ants', 'bees', 'ants']](../_images/sphx_glr_transfer_learning_tutorial_001.png)
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]
92%|█████████▏| 40.9M/44.7M [00:00<00:00, 428MB/s]
100%|██████████| 44.7M/44.7M [00:00<00:00, 429MB/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: 0.7158 Acc: 0.6352
val Loss: 0.1669 Acc: 0.9542
Epoch 1/24
----------
train Loss: 0.4513 Acc: 0.8074
val Loss: 0.2216 Acc: 0.9085
Epoch 2/24
----------
train Loss: 0.5663 Acc: 0.7869
val Loss: 0.2755 Acc: 0.9281
Epoch 3/24
----------
train Loss: 0.4882 Acc: 0.7992
val Loss: 0.3038 Acc: 0.8824
Epoch 4/24
----------
train Loss: 0.4776 Acc: 0.7992
val Loss: 0.3273 Acc: 0.8627
Epoch 5/24
----------
train Loss: 0.4920 Acc: 0.7992
val Loss: 0.2407 Acc: 0.9346
Epoch 6/24
----------
train Loss: 0.5417 Acc: 0.8156
val Loss: 0.4268 Acc: 0.8627
Epoch 7/24
----------
train Loss: 0.3231 Acc: 0.8607
val Loss: 0.2066 Acc: 0.9020
Epoch 8/24
----------
train Loss: 0.3396 Acc: 0.8443
val Loss: 0.2411 Acc: 0.9150
Epoch 9/24
----------
train Loss: 0.4098 Acc: 0.8443
val Loss: 0.2532 Acc: 0.9020
Epoch 10/24
----------
train Loss: 0.3669 Acc: 0.8484
val Loss: 0.2408 Acc: 0.9085
Epoch 11/24
----------
train Loss: 0.3095 Acc: 0.8607
val Loss: 0.2160 Acc: 0.9216
Epoch 12/24
----------
train Loss: 0.2510 Acc: 0.9098
val Loss: 0.2337 Acc: 0.9020
Epoch 13/24
----------
train Loss: 0.2760 Acc: 0.8648
val Loss: 0.2547 Acc: 0.9085
Epoch 14/24
----------
train Loss: 0.3286 Acc: 0.8607
val Loss: 0.2774 Acc: 0.8889
Epoch 15/24
----------
train Loss: 0.2908 Acc: 0.8648
val Loss: 0.2383 Acc: 0.9085
Epoch 16/24
----------
train Loss: 0.2483 Acc: 0.8770
val Loss: 0.2408 Acc: 0.9150
Epoch 17/24
----------
train Loss: 0.2537 Acc: 0.9016
val Loss: 0.2317 Acc: 0.9150
Epoch 18/24
----------
train Loss: 0.2198 Acc: 0.8975
val Loss: 0.2328 Acc: 0.9085
Epoch 19/24
----------
train Loss: 0.2642 Acc: 0.8811
val Loss: 0.2155 Acc: 0.9085
Epoch 20/24
----------
train Loss: 0.3511 Acc: 0.8648
val Loss: 0.2252 Acc: 0.9150
Epoch 21/24
----------
train Loss: 0.2156 Acc: 0.9057
val Loss: 0.2755 Acc: 0.8889
Epoch 22/24
----------
train Loss: 0.2645 Acc: 0.8934
val Loss: 0.2253 Acc: 0.9085
Epoch 23/24
----------
train Loss: 0.2238 Acc: 0.8934
val Loss: 0.2474 Acc: 0.9085
Epoch 24/24
----------
train Loss: 0.2527 Acc: 0.8893
val Loss: 0.2524 Acc: 0.9020
Training complete in 0m 34s
Best val Acc: 0.954248
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: 0.6188 Acc: 0.6352
val Loss: 0.2699 Acc: 0.9020
Epoch 1/24
----------
train Loss: 0.6212 Acc: 0.7377
val Loss: 0.2149 Acc: 0.9412
Epoch 2/24
----------
train Loss: 0.4927 Acc: 0.7828
val Loss: 0.2268 Acc: 0.9281
Epoch 3/24
----------
train Loss: 0.4410 Acc: 0.7992
val Loss: 0.1737 Acc: 0.9542
Epoch 4/24
----------
train Loss: 0.4862 Acc: 0.8074
val Loss: 0.1891 Acc: 0.9542
Epoch 5/24
----------
train Loss: 0.5222 Acc: 0.8115
val Loss: 0.1829 Acc: 0.9346
Epoch 6/24
----------
train Loss: 0.3929 Acc: 0.8238
val Loss: 0.3061 Acc: 0.9085
Epoch 7/24
----------
train Loss: 0.3715 Acc: 0.8279
val Loss: 0.1858 Acc: 0.9412
Epoch 8/24
----------
train Loss: 0.3504 Acc: 0.8361
val Loss: 0.1750 Acc: 0.9542
Epoch 9/24
----------
train Loss: 0.3893 Acc: 0.8156
val Loss: 0.1880 Acc: 0.9477
Epoch 10/24
----------
train Loss: 0.3844 Acc: 0.8279
val Loss: 0.1963 Acc: 0.9412
Epoch 11/24
----------
train Loss: 0.3145 Acc: 0.8525
val Loss: 0.1951 Acc: 0.9477
Epoch 12/24
----------
train Loss: 0.3444 Acc: 0.8320
val Loss: 0.1808 Acc: 0.9477
Epoch 13/24
----------
train Loss: 0.3654 Acc: 0.8279
val Loss: 0.1754 Acc: 0.9477
Epoch 14/24
----------
train Loss: 0.3119 Acc: 0.8484
val Loss: 0.2020 Acc: 0.9412
Epoch 15/24
----------
train Loss: 0.3279 Acc: 0.8607
val Loss: 0.1886 Acc: 0.9412
Epoch 16/24
----------
train Loss: 0.2890 Acc: 0.8811
val Loss: 0.2007 Acc: 0.9412
Epoch 17/24
----------
train Loss: 0.3300 Acc: 0.8484
val Loss: 0.1944 Acc: 0.9412
Epoch 18/24
----------
train Loss: 0.3364 Acc: 0.8402
val Loss: 0.2108 Acc: 0.9346
Epoch 19/24
----------
train Loss: 0.4361 Acc: 0.7951
val Loss: 0.1980 Acc: 0.9412
Epoch 20/24
----------
train Loss: 0.2946 Acc: 0.8730
val Loss: 0.1981 Acc: 0.9477
Epoch 21/24
----------
train Loss: 0.2723 Acc: 0.8730
val Loss: 0.1797 Acc: 0.9477
Epoch 22/24
----------
train Loss: 0.3166 Acc: 0.8730
val Loss: 0.1878 Acc: 0.9412
Epoch 23/24
----------
train Loss: 0.2998 Acc: 0.8730
val Loss: 0.1825 Acc: 0.9412
Epoch 24/24
----------
train Loss: 0.3583 Acc: 0.8525
val Loss: 0.1824 Acc: 0.9477
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.
Total running time of the script: (1 minutes 3.479 seconds)
