注意
单击此处下载完整的示例代码
Spatial Transformer 网络教程¶
创建时间： 2017-11-08 |上次更新时间：2024 年 1 月 19 日 |上次验证： Nov 05， 2024
作者： Ghassen HAMROUNI
在本教程中，您将学习如何使用 一种称为空间转换器的视觉注意机制 网络。您可以阅读有关 spatial transformer 的更多信息 DeepMind 论文中的网络
空间变换器网络是 Differentiable 的泛化 注意任何空间变换。空间变换器网络 （简称 STN）允许神经网络学习如何执行空间 对输入图像进行变换，以增强几何图形 模型的不变性。 例如，它可以裁剪感兴趣的区域、缩放和校正 图像的方向。这可能是一种有用的机制，因为 CNN 对 rotation 和 scale 以及更通用的仿射 转换。
STN 的一大优点是能够简单地将其插入 任何现有的 CNN，只需稍作修改即可。
# 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 0x7f6a304bde10>
加载数据¶
在这篇文章中，我们将试验经典的 MNIST 数据集。使用 使用空间转换器增强的标准卷积网络 网络。
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)
Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz
Failed to download (trying next):
HTTP Error 403: Forbidden
Downloading https://ossci-datasets.s3.amazonaws.com/mnist/train-images-idx3-ubyte.gz
Downloading https://ossci-datasets.s3.amazonaws.com/mnist/train-images-idx3-ubyte.gz to ./MNIST/raw/train-images-idx3-ubyte.gz
0%| | 0.00/9.91M [00:00<?, ?B/s]
100%|##########| 9.91M/9.91M [00:00<00:00, 131MB/s]
Extracting ./MNIST/raw/train-images-idx3-ubyte.gz to ./MNIST/raw
Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz
Failed to download (trying next):
HTTP Error 403: Forbidden
Downloading https://ossci-datasets.s3.amazonaws.com/mnist/train-labels-idx1-ubyte.gz
Downloading https://ossci-datasets.s3.amazonaws.com/mnist/train-labels-idx1-ubyte.gz to ./MNIST/raw/train-labels-idx1-ubyte.gz
0%| | 0.00/28.9k [00:00<?, ?B/s]
100%|##########| 28.9k/28.9k [00:00<00:00, 18.6MB/s]
Extracting ./MNIST/raw/train-labels-idx1-ubyte.gz to ./MNIST/raw
Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz
Failed to download (trying next):
HTTP Error 403: Forbidden
Downloading https://ossci-datasets.s3.amazonaws.com/mnist/t10k-images-idx3-ubyte.gz
Downloading https://ossci-datasets.s3.amazonaws.com/mnist/t10k-images-idx3-ubyte.gz to ./MNIST/raw/t10k-images-idx3-ubyte.gz
0%| | 0.00/1.65M [00:00<?, ?B/s]
100%|##########| 1.65M/1.65M [00:00<00:00, 206MB/s]
Extracting ./MNIST/raw/t10k-images-idx3-ubyte.gz to ./MNIST/raw
Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz
Failed to download (trying next):
HTTP Error 403: Forbidden
Downloading https://ossci-datasets.s3.amazonaws.com/mnist/t10k-labels-idx1-ubyte.gz
Downloading https://ossci-datasets.s3.amazonaws.com/mnist/t10k-labels-idx1-ubyte.gz to ./MNIST/raw/t10k-labels-idx1-ubyte.gz
0%| | 0.00/4.54k [00:00<?, ?B/s]
100%|##########| 4.54k/4.54k [00:00<00:00, 23.5MB/s]
Extracting ./MNIST/raw/t10k-labels-idx1-ubyte.gz to ./MNIST/raw
描绘空间变压器网络¶
空间变换器网络归结为三个主要组成部分：
本地化网络是一个常规的 CNN，它将 转换参数。转变从来没有被学习过 相反，网络会自动从这个数据集中学习 提高全局精度的空间变换。
网格生成器在输入中生成坐标网格 与输出图像中的每个像素对应的图像。
采样器使用转换的参数并应用 it 添加到输入图像中。
注意
我们需要最新版本的 PyTorch，其中包含 affine_grid 和 grid_sample 模块。
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)
训练模型¶
现在，让我们使用 SGD 算法来训练模型。网络是 以监督的方式学习分类任务。同时 该模型以端到端方式自动学习 STN。
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)))
可视化 STN 结果¶
现在，我们将检查我们习得的视觉注意的结果 机制。
我们定义了一个小的辅助函数，以便可视化 训练时的转变。
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()
/usr/local/lib/python3.10/dist-packages/torch/nn/functional.py:4969: 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.
/usr/local/lib/python3.10/dist-packages/torch/nn/functional.py:4902: 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.
Train Epoch: 1 [0/60000 (0%)] Loss: 2.315648
Train Epoch: 1 [32000/60000 (53%)] Loss: 1.056807
/usr/local/lib/python3.10/dist-packages/torch/nn/_reduction.py:51: UserWarning:
size_average and reduce args will be deprecated, please use reduction='sum' instead.
Test set: Average loss: 0.2980, Accuracy: 9146/10000 (91%)
Train Epoch: 2 [0/60000 (0%)] Loss: 0.576728
Train Epoch: 2 [32000/60000 (53%)] Loss: 0.357588
Test set: Average loss: 0.1712, Accuracy: 9480/10000 (95%)
Train Epoch: 3 [0/60000 (0%)] Loss: 0.375578
Train Epoch: 3 [32000/60000 (53%)] Loss: 0.215431
Test set: Average loss: 0.1544, Accuracy: 9520/10000 (95%)
Train Epoch: 4 [0/60000 (0%)] Loss: 0.401764
Train Epoch: 4 [32000/60000 (53%)] Loss: 0.112544
Test set: Average loss: 0.1233, Accuracy: 9629/10000 (96%)
Train Epoch: 5 [0/60000 (0%)] Loss: 0.296059
Train Epoch: 5 [32000/60000 (53%)] Loss: 0.203347
Test set: Average loss: 0.0925, Accuracy: 9719/10000 (97%)
Train Epoch: 6 [0/60000 (0%)] Loss: 0.107811
Train Epoch: 6 [32000/60000 (53%)] Loss: 0.114046
Test set: Average loss: 0.0823, Accuracy: 9765/10000 (98%)
Train Epoch: 7 [0/60000 (0%)] Loss: 0.130760
Train Epoch: 7 [32000/60000 (53%)] Loss: 0.154117
Test set: Average loss: 0.0742, Accuracy: 9778/10000 (98%)
Train Epoch: 8 [0/60000 (0%)] Loss: 0.270492
Train Epoch: 8 [32000/60000 (53%)] Loss: 0.125196
Test set: Average loss: 0.0639, Accuracy: 9804/10000 (98%)
Train Epoch: 9 [0/60000 (0%)] Loss: 0.116596
Train Epoch: 9 [32000/60000 (53%)] Loss: 0.139002
Test set: Average loss: 0.0596, Accuracy: 9817/10000 (98%)
Train Epoch: 10 [0/60000 (0%)] Loss: 0.076979
Train Epoch: 10 [32000/60000 (53%)] Loss: 0.176808
Test set: Average loss: 0.0682, Accuracy: 9795/10000 (98%)
Train Epoch: 11 [0/60000 (0%)] Loss: 0.154275
Train Epoch: 11 [32000/60000 (53%)] Loss: 0.151468
Test set: Average loss: 0.0646, Accuracy: 9807/10000 (98%)
Train Epoch: 12 [0/60000 (0%)] Loss: 0.130388
Train Epoch: 12 [32000/60000 (53%)] Loss: 0.116185
Test set: Average loss: 0.0527, Accuracy: 9846/10000 (98%)
Train Epoch: 13 [0/60000 (0%)] Loss: 0.131821
Train Epoch: 13 [32000/60000 (53%)] Loss: 0.134187
Test set: Average loss: 0.0657, Accuracy: 9806/10000 (98%)
Train Epoch: 14 [0/60000 (0%)] Loss: 0.066973
Train Epoch: 14 [32000/60000 (53%)] Loss: 0.162483
Test set: Average loss: 0.0493, Accuracy: 9851/10000 (99%)
Train Epoch: 15 [0/60000 (0%)] Loss: 0.043320
Train Epoch: 15 [32000/60000 (53%)] Loss: 0.063544
Test set: Average loss: 0.0495, Accuracy: 9850/10000 (98%)
Train Epoch: 16 [0/60000 (0%)] Loss: 0.086180
Train Epoch: 16 [32000/60000 (53%)] Loss: 0.181362
Test set: Average loss: 0.0468, Accuracy: 9866/10000 (99%)
Train Epoch: 17 [0/60000 (0%)] Loss: 0.285877
Train Epoch: 17 [32000/60000 (53%)] Loss: 0.100507
Test set: Average loss: 0.0467, Accuracy: 9851/10000 (99%)
Train Epoch: 18 [0/60000 (0%)] Loss: 0.041245
Train Epoch: 18 [32000/60000 (53%)] Loss: 0.124705
Test set: Average loss: 0.0453, Accuracy: 9871/10000 (99%)
Train Epoch: 19 [0/60000 (0%)] Loss: 0.041999
Train Epoch: 19 [32000/60000 (53%)] Loss: 0.148507
Test set: Average loss: 0.0457, Accuracy: 9867/10000 (99%)
Train Epoch: 20 [0/60000 (0%)] Loss: 0.054914
Train Epoch: 20 [32000/60000 (53%)] Loss: 0.017267
Test set: Average loss: 0.0432, Accuracy: 9881/10000 (99%)
脚本总运行时间：（2 分 17.768 秒）