空间变压器网络教程 ¶
译者:片刻小哥哥
项目地址:https://pytorch.apachecn.org/2.0/tutorials/intermediate/spatial_transformer_tutorial
原始地址:https://pytorch.org/tutorials/intermediate/spatial_transformer_tutorial.html
作者 : Ghassen HAMROUNI
在本教程中,您将学习如何使用称为空间变换器网络的视觉注意机制来增强网络。您可以在 DeepMind 论文 中阅读有关空间变换器 网络的更多信息
空间变换网络是对任何空间变换的可微分注意力的概括。空间变换网络 (简称 STN)允许神经网络学习如何对输入图像执行空间 变换,以增强模型的几何 方差。 例如,它可以裁剪感兴趣的区域,缩放并纠正图像的方向。它可能是一种有用的机制,因为 CNN 对于旋转和缩放以及更一般的仿射变换 不是不变的。
STN 最好的事情之一是能够简单地将其插入 任何现有的 CNN,只需很少的修改。
# 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
正在加载数据 ¶
在这篇文章中,我们使用经典的 MNIST 数据集进行实验。使用 通过空间变换器 网络增强的标准卷积网络。
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
Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz to ./MNIST/raw/train-images-idx3-ubyte.gz
0%| | 0/9912422 [00:00<?, ?it/s]
100%|##########| 9912422/9912422 [00:00<00:00, 322702592.79it/s]
Extracting ./MNIST/raw/train-images-idx3-ubyte.gz to ./MNIST/raw
Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz
Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz to ./MNIST/raw/train-labels-idx1-ubyte.gz
0%| | 0/28881 [00:00<?, ?it/s]
100%|##########| 28881/28881 [00:00<00:00, 46770538.16it/s]
Extracting ./MNIST/raw/train-labels-idx1-ubyte.gz to ./MNIST/raw
Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz
Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz to ./MNIST/raw/t10k-images-idx3-ubyte.gz
0%| | 0/1648877 [00:00<?, ?it/s]
100%|##########| 1648877/1648877 [00:00<00:00, 162999160.87it/s]
Extracting ./MNIST/raw/t10k-images-idx3-ubyte.gz to ./MNIST/raw
Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz
Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz to ./MNIST/raw/t10k-labels-idx1-ubyte.gz
0%| | 0/4542 [00:00<?, ?it/s]
100%|##########| 4542/4542 [00:00<00:00, 34202026.51it/s]
Extracting ./MNIST/raw/t10k-labels-idx1-ubyte.gz to ./MNIST/raw
描述空间变换器网络 ¶
空间变换器网络可归结为三个主要组件:
- 定位网络是一个常规的 CNN,它对变换参数进行回归。从未从该数据集中 显式学习变换,而是网络自动学习 提高全局精度的空间变换。
- 网格生成器在输入图像中生成 对应于输出图像中每个像素的坐标网格.
- 采样器使用变换参数并将其应用于 输入图像。
注意
我们需要包含 affine_grid 和 grid_sample 模块的最新版本的 PyTorch。
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’s使用SGD算法来训练模型。网络正在以监督方式学习分类任务。同时 模型以端到端的方式自动学习 STN。
optimizer = optim.SGD(model.parameters(), lr=0.01)
def train(epoch):
[model.train](https://pytorch.org/docs/stable/generated/torch.nn.Module.html#torch.nn.Module.train "torch.nn.Module. train")()
for batch_idx, (data, target) in enumerate([train_loader](https://pytorch.org/docs/stable/data.html#torch.utils.data. DataLoader "torch.utils.data.DataLoader")):
data, target = data.to([device](https://pytorch.org/docs/stable/tensor_attributes.html#torch.device "torch.device ")), target.to([device](https://pytorch.org/docs/stable/tensor_attributes.html#torch.device "torch.device"))
[optimizer.zero_grad] (https://pytorch.org/docs/stable/generated/torch.optim.SGD.html#torch.optim.SGD.zero_grad“torch.optim.SGD.zero_grad”)()
输出=模型(数据)
损失 = [F.nll_loss](https://pytorch.org/docs/stable/generated/torch.nn.function.nll_loss.html#torch.nn.function.nll_loss "torch.nn.function.nll_loss")(输出,目标)
loss.backward()
[optimizer.step](https://pytorch.org/docs/stable/generated/torch.optim.SGD.html#torch.optim。 SGD.step "torch.optim.SGD.step")()
if batch_idx % 500 == 0:
print('训练纪元: {} [{}/{} ({:.0f }%)]\tLoss: {:.6f}'.format(
epoch, batch_idx * len(data), len([train_loader.dataset](https://pytorch.org/vision/stable/generated/torchvision.datasets.MNIST.html#torchvision.datasets.MNIST "torchvision.datasets.MNIST")),
100. * batch_idx /len([train\ _loader](https://pytorch.org/docs/stable/data.html#torch.utils.data.DataLoader "torch.utils.data.DataLoader")), loss.item()))
#
# 一个简单的测试程序来测量 STN 在 MNIST 上的性能。
#
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('Test set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)'
.format(test_loss, correct, len(test_loader.dataset),
100. * correct / len(test_loader.dataset)))
可视化 STN 结果 ¶
现在,我们将检查学习到的视觉注意力 机制的结果。
我们定义了一个小辅助函数,以便在训练时 可视化 转换。
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()
/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/nn/functional.py:4358: 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.
/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/nn/functional.py:4296: 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: 1.047744
/opt/conda/envs/py_3.10/lib/python3.10/site-packages/torch/nn/_reduction.py:42: UserWarning:
size_average and reduce args will be deprecated, please use reduction='sum' instead.
Test set: Average loss: 0.2656, Accuracy: 9264/10000 (93%)
Train Epoch: 2 [0/60000 (0%)] Loss: 0.533303
Train Epoch: 2 [32000/60000 (53%)] Loss: 0.331733
Test set: Average loss: 0.1831, Accuracy: 9462/10000 (95%)
Train Epoch: 3 [0/60000 (0%)] Loss: 0.387849
Train Epoch: 3 [32000/60000 (53%)] Loss: 0.215252
Test set: Average loss: 0.1148, Accuracy: 9656/10000 (97%)
Train Epoch: 4 [0/60000 (0%)] Loss: 0.338932
Train Epoch: 4 [32000/60000 (53%)] Loss: 0.213857
Test set: Average loss: 0.1616, Accuracy: 9491/10000 (95%)
Train Epoch: 5 [0/60000 (0%)] Loss: 0.305876
Train Epoch: 5 [32000/60000 (53%)] Loss: 0.217289
Test set: Average loss: 0.1351, Accuracy: 9609/10000 (96%)
Train Epoch: 6 [0/60000 (0%)] Loss: 0.221662
Train Epoch: 6 [32000/60000 (53%)] Loss: 0.145264
Test set: Average loss: 0.0708, Accuracy: 9782/10000 (98%)
Train Epoch: 7 [0/60000 (0%)] Loss: 0.114100
Train Epoch: 7 [32000/60000 (53%)] Loss: 0.190583
Test set: Average loss: 0.0742, Accuracy: 9766/10000 (98%)
Train Epoch: 8 [0/60000 (0%)] Loss: 0.293466
Train Epoch: 8 [32000/60000 (53%)] Loss: 0.070622
Test set: Average loss: 0.0616, Accuracy: 9821/10000 (98%)
Train Epoch: 9 [0/60000 (0%)] Loss: 0.092730
Train Epoch: 9 [32000/60000 (53%)] Loss: 0.080178
Test set: Average loss: 0.0776, Accuracy: 9766/10000 (98%)
Train Epoch: 10 [0/60000 (0%)] Loss: 0.095328
Train Epoch: 10 [32000/60000 (53%)] Loss: 0.227478
Test set: Average loss: 0.0738, Accuracy: 9776/10000 (98%)
Train Epoch: 11 [0/60000 (0%)] Loss: 0.162715
Train Epoch: 11 [32000/60000 (53%)] Loss: 0.114082
Test set: Average loss: 0.0616, Accuracy: 9810/10000 (98%)
Train Epoch: 12 [0/60000 (0%)] Loss: 0.105073
Train Epoch: 12 [32000/60000 (53%)] Loss: 0.184882
Test set: Average loss: 0.0538, Accuracy: 9839/10000 (98%)
Train Epoch: 13 [0/60000 (0%)] Loss: 0.129685
Train Epoch: 13 [32000/60000 (53%)] Loss: 0.138069
Test set: Average loss: 0.0522, Accuracy: 9838/10000 (98%)
Train Epoch: 14 [0/60000 (0%)] Loss: 0.046923
Train Epoch: 14 [32000/60000 (53%)] Loss: 0.100477
Test set: Average loss: 0.0514, Accuracy: 9849/10000 (98%)
Train Epoch: 15 [0/60000 (0%)] Loss: 0.063011
Train Epoch: 15 [32000/60000 (53%)] Loss: 0.158940
Test set: Average loss: 0.0734, Accuracy: 9765/10000 (98%)
Train Epoch: 16 [0/60000 (0%)] Loss: 0.076108
Train Epoch: 16 [32000/60000 (53%)] Loss: 0.149375
Test set: Average loss: 0.0452, Accuracy: 9857/10000 (99%)
Train Epoch: 17 [0/60000 (0%)] Loss: 0.266226
Train Epoch: 17 [32000/60000 (53%)] Loss: 0.184768
Test set: Average loss: 0.0857, Accuracy: 9746/10000 (97%)
Train Epoch: 18 [0/60000 (0%)] Loss: 0.112116
Train Epoch: 18 [32000/60000 (53%)] Loss: 0.089787
Test set: Average loss: 0.0583, Accuracy: 9833/10000 (98%)
Train Epoch: 19 [0/60000 (0%)] Loss: 0.065648
Train Epoch: 19 [32000/60000 (53%)] Loss: 0.143108
Test set: Average loss: 0.0442, Accuracy: 9863/10000 (99%)
Train Epoch: 20 [0/60000 (0%)] Loss: 0.071892
Train Epoch: 20 [32000/60000 (53%)] Loss: 0.154807
Test set: Average loss: 0.0489, Accuracy: 9851/10000 (99%)
脚本的总运行时间: ( 2 分 4.495 秒)