Note
Go to the end to download the full example code.
Spatial Transformer Networks Tutorial#
Created On: Nov 08, 2017 | Last Updated: Jan 19, 2024 | Last Verified: Nov 05, 2024
Author: Ghassen HAMROUNI
In this tutorial, you will learn how to augment your network using a visual attention mechanism called spatial transformer networks. You can read more about the spatial transformer networks in the DeepMind paper
Spatial transformer networks are a generalization of differentiable attention to any spatial transformation. Spatial transformer networks (STN for short) allow a neural network to learn how to perform spatial transformations on the input image in order to enhance the geometric invariance of the model. For example, it can crop a region of interest, scale and correct the orientation of an image. It can be a useful mechanism because CNNs are not invariant to rotation and scale and more general affine transformations.
One of the best things about STN is the ability to simply plug it into any existing CNN with very little modification.
# License: BSD
# Author: Ghassen Hamrouni
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
import torchvision
from torchvision import datasets, transforms
import matplotlib.pyplot as plt
import numpy as np
plt.ion() # interactive mode
<contextlib.ExitStack object at 0x7f97a234eaa0>
Loading the data#
In this post we experiment with the classic MNIST dataset. Using a standard convolutional network augmented with a spatial transformer network.
from six.moves import urllib
opener = urllib.request.build_opener()
opener.addheaders = [('User-agent', 'Mozilla/5.0')]
urllib.request.install_opener(opener)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# Training dataset
train_loader = torch.utils.data.DataLoader(
datasets.MNIST(root='.', train=True, download=True,
transform=transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.1307,), (0.3081,))
])), batch_size=64, shuffle=True, num_workers=4)
# Test dataset
test_loader = torch.utils.data.DataLoader(
datasets.MNIST(root='.', train=False, transform=transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.1307,), (0.3081,))
])), batch_size=64, shuffle=True, num_workers=4)
0%| | 0.00/9.91M [00:00<?, ?B/s]
100%|██████████| 9.91M/9.91M [00:00<00:00, 151MB/s]
0%| | 0.00/28.9k [00:00<?, ?B/s]
100%|██████████| 28.9k/28.9k [00:00<00:00, 47.6MB/s]
0%| | 0.00/1.65M [00:00<?, ?B/s]
100%|██████████| 1.65M/1.65M [00:00<00:00, 139MB/s]
0%| | 0.00/4.54k [00:00<?, ?B/s]
100%|██████████| 4.54k/4.54k [00:00<00:00, 28.0MB/s]
Depicting spatial transformer networks#
Spatial transformer networks boils down to three main components :
The localization network is a regular CNN which regresses the transformation parameters. The transformation is never learned explicitly from this dataset, instead the network learns automatically the spatial transformations that enhances the global accuracy.
The grid generator generates a grid of coordinates in the input image corresponding to each pixel from the output image.
The sampler uses the parameters of the transformation and applies it to the input image.
Note
We need the latest version of PyTorch that contains affine_grid and grid_sample modules.
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
self.conv2_drop = nn.Dropout2d()
self.fc1 = nn.Linear(320, 50)
self.fc2 = nn.Linear(50, 10)
# Spatial transformer localization-network
self.localization = nn.Sequential(
nn.Conv2d(1, 8, kernel_size=7),
nn.MaxPool2d(2, stride=2),
nn.ReLU(True),
nn.Conv2d(8, 10, kernel_size=5),
nn.MaxPool2d(2, stride=2),
nn.ReLU(True)
)
# Regressor for the 3 * 2 affine matrix
self.fc_loc = nn.Sequential(
nn.Linear(10 * 3 * 3, 32),
nn.ReLU(True),
nn.Linear(32, 3 * 2)
)
# Initialize the weights/bias with identity transformation
self.fc_loc[2].weight.data.zero_()
self.fc_loc[2].bias.data.copy_(torch.tensor([1, 0, 0, 0, 1, 0], dtype=torch.float))
# Spatial transformer network forward function
def stn(self, x):
xs = self.localization(x)
xs = xs.view(-1, 10 * 3 * 3)
theta = self.fc_loc(xs)
theta = theta.view(-1, 2, 3)
grid = F.affine_grid(theta, x.size())
x = F.grid_sample(x, grid)
return x
def forward(self, x):
# transform the input
x = self.stn(x)
# Perform the usual forward pass
x = F.relu(F.max_pool2d(self.conv1(x), 2))
x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))
x = x.view(-1, 320)
x = F.relu(self.fc1(x))
x = F.dropout(x, training=self.training)
x = self.fc2(x)
return F.log_softmax(x, dim=1)
model = Net().to(device)
Training the model#
Now, let’s use the SGD algorithm to train the model. The network is learning the classification task in a supervised way. In the same time the model is learning STN automatically in an end-to-end fashion.
optimizer = optim.SGD(model.parameters(), lr=0.01)
def train(epoch):
model.train()
for batch_idx, (data, target) in enumerate(train_loader):
data, target = data.to(device), target.to(device)
optimizer.zero_grad()
output = model(data)
loss = F.nll_loss(output, target)
loss.backward()
optimizer.step()
if batch_idx % 500 == 0:
print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
epoch, batch_idx * len(data), len(train_loader.dataset),
100. * batch_idx / len(train_loader), loss.item()))
#
# A simple test procedure to measure the STN performances on MNIST.
#
def test():
with torch.no_grad():
model.eval()
test_loss = 0
correct = 0
for data, target in test_loader:
data, target = data.to(device), target.to(device)
output = model(data)
# sum up batch loss
test_loss += F.nll_loss(output, target, size_average=False).item()
# get the index of the max log-probability
pred = output.max(1, keepdim=True)[1]
correct += pred.eq(target.view_as(pred)).sum().item()
test_loss /= len(test_loader.dataset)
print('\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\n'
.format(test_loss, correct, len(test_loader.dataset),
100. * correct / len(test_loader.dataset)))
Visualizing the STN results#
Now, we will inspect the results of our learned visual attention mechanism.
We define a small helper function in order to visualize the transformations while training.
def convert_image_np(inp):
"""Convert a Tensor to numpy image."""
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)
return inp
# We want to visualize the output of the spatial transformers layer
# after the training, we visualize a batch of input images and
# the corresponding transformed batch using STN.
def visualize_stn():
with torch.no_grad():
# Get a batch of training data
data = next(iter(test_loader))[0].to(device)
input_tensor = data.cpu()
transformed_input_tensor = model.stn(data).cpu()
in_grid = convert_image_np(
torchvision.utils.make_grid(input_tensor))
out_grid = convert_image_np(
torchvision.utils.make_grid(transformed_input_tensor))
# Plot the results side-by-side
f, axarr = plt.subplots(1, 2)
axarr[0].imshow(in_grid)
axarr[0].set_title('Dataset Images')
axarr[1].imshow(out_grid)
axarr[1].set_title('Transformed Images')
for epoch in range(1, 20 + 1):
train(epoch)
test()
# Visualize the STN transformation on some input batch
visualize_stn()
plt.ioff()
plt.show()
/var/lib/workspace/intermediate_source/spatial_transformer_tutorial.py:130: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details.
grid = F.affine_grid(theta, x.size())
/var/lib/workspace/intermediate_source/spatial_transformer_tutorial.py:131: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details.
x = F.grid_sample(x, grid)
Train Epoch: 1 [0/60000 (0%)] Loss: 2.323574
Train Epoch: 1 [32000/60000 (53%)] Loss: 0.589004
/usr/local/lib/python3.10/dist-packages/torch/nn/functional.py:3178: UserWarning: size_average and reduce args will be deprecated, please use reduction='sum' instead.
reduction = _Reduction.legacy_get_string(size_average, reduce)
Test set: Average loss: 0.2863, Accuracy: 9142/10000 (91%)
Train Epoch: 2 [0/60000 (0%)] Loss: 0.677977
Train Epoch: 2 [32000/60000 (53%)] Loss: 0.430223
Test set: Average loss: 0.1350, Accuracy: 9611/10000 (96%)
Train Epoch: 3 [0/60000 (0%)] Loss: 0.304662
Train Epoch: 3 [32000/60000 (53%)] Loss: 0.158780
Test set: Average loss: 0.0888, Accuracy: 9712/10000 (97%)
Train Epoch: 4 [0/60000 (0%)] Loss: 0.260996
Train Epoch: 4 [32000/60000 (53%)] Loss: 0.287503
Test set: Average loss: 0.0876, Accuracy: 9744/10000 (97%)
Train Epoch: 5 [0/60000 (0%)] Loss: 0.194340
Train Epoch: 5 [32000/60000 (53%)] Loss: 0.247315
Test set: Average loss: 0.0783, Accuracy: 9750/10000 (98%)
Train Epoch: 6 [0/60000 (0%)] Loss: 0.213213
Train Epoch: 6 [32000/60000 (53%)] Loss: 0.154125
Test set: Average loss: 0.0592, Accuracy: 9821/10000 (98%)
Train Epoch: 7 [0/60000 (0%)] Loss: 0.193532
Train Epoch: 7 [32000/60000 (53%)] Loss: 0.206896
Test set: Average loss: 0.0589, Accuracy: 9826/10000 (98%)
Train Epoch: 8 [0/60000 (0%)] Loss: 0.051803
Train Epoch: 8 [32000/60000 (53%)] Loss: 0.115960
Test set: Average loss: 0.0540, Accuracy: 9836/10000 (98%)
Train Epoch: 9 [0/60000 (0%)] Loss: 0.165460
Train Epoch: 9 [32000/60000 (53%)] Loss: 0.142512
Test set: Average loss: 0.0498, Accuracy: 9846/10000 (98%)
Train Epoch: 10 [0/60000 (0%)] Loss: 0.162217
Train Epoch: 10 [32000/60000 (53%)] Loss: 0.142307
Test set: Average loss: 0.0459, Accuracy: 9853/10000 (99%)
Train Epoch: 11 [0/60000 (0%)] Loss: 0.051161
Train Epoch: 11 [32000/60000 (53%)] Loss: 0.202848
Test set: Average loss: 0.0485, Accuracy: 9850/10000 (98%)
Train Epoch: 12 [0/60000 (0%)] Loss: 0.168068
Train Epoch: 12 [32000/60000 (53%)] Loss: 0.058105
Test set: Average loss: 0.0492, Accuracy: 9851/10000 (99%)
Train Epoch: 13 [0/60000 (0%)] Loss: 0.106750
Train Epoch: 13 [32000/60000 (53%)] Loss: 0.113328
Test set: Average loss: 0.0517, Accuracy: 9864/10000 (99%)
Train Epoch: 14 [0/60000 (0%)] Loss: 0.149636
Train Epoch: 14 [32000/60000 (53%)] Loss: 0.116727
Test set: Average loss: 0.0403, Accuracy: 9864/10000 (99%)
Train Epoch: 15 [0/60000 (0%)] Loss: 0.043748
Train Epoch: 15 [32000/60000 (53%)] Loss: 0.093911
Test set: Average loss: 0.0610, Accuracy: 9827/10000 (98%)
Train Epoch: 16 [0/60000 (0%)] Loss: 0.074828
Train Epoch: 16 [32000/60000 (53%)] Loss: 0.027772
Test set: Average loss: 0.0393, Accuracy: 9881/10000 (99%)
Train Epoch: 17 [0/60000 (0%)] Loss: 0.028084
Train Epoch: 17 [32000/60000 (53%)] Loss: 0.017267
Test set: Average loss: 0.0382, Accuracy: 9885/10000 (99%)
Train Epoch: 18 [0/60000 (0%)] Loss: 0.156877
Train Epoch: 18 [32000/60000 (53%)] Loss: 0.134994
Test set: Average loss: 0.0704, Accuracy: 9773/10000 (98%)
Train Epoch: 19 [0/60000 (0%)] Loss: 0.168051
Train Epoch: 19 [32000/60000 (53%)] Loss: 0.185395
Test set: Average loss: 0.0384, Accuracy: 9887/10000 (99%)
Train Epoch: 20 [0/60000 (0%)] Loss: 0.047536
Train Epoch: 20 [32000/60000 (53%)] Loss: 0.016828
Test set: Average loss: 0.0358, Accuracy: 9896/10000 (99%)
Total running time of the script: (1 minutes 38.609 seconds)
