服务器之家

服务器之家 > 正文

使用PyTorch实现MNIST手写体识别代码

时间:2020-04-14 10:31     来源/作者:kaijie234

实验环境

win10 + anaconda + jupyter notebook

Pytorch1.1.0

Python3.7

gpu环境(可选)

MNIST数据集介绍

MNIST 包括6万张28x28的训练样本,1万张测试样本,可以说是CV里的“Hello Word”。本文使用的CNN网络将MNIST数据的识别率提高到了99%。下面我们就开始进行实战。

导入包

?
1
2
3
4
5
6
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torchvision import datasets, transforms
torch.__version__

定义超参数

?
1
2
3
BATCH_SIZE=512
EPOCHS=20
DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")

数据集

我们直接使用PyTorch中自带的dataset,并使用DataLoader对训练数据和测试数据分别进行读取。如果下载过数据集这里download可选择False

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
train_loader = torch.utils.data.DataLoader(
    datasets.MNIST('data', train=True, download=True,
            transform=transforms.Compose([
              transforms.ToTensor(),
              transforms.Normalize((0.1307,), (0.3081,))
            ])),
    batch_size=BATCH_SIZE, shuffle=True)
 
test_loader = torch.utils.data.DataLoader(
    datasets.MNIST('data', train=False, transform=transforms.Compose([
              transforms.ToTensor(),
              transforms.Normalize((0.1307,), (0.3081,))
            ])),
    batch_size=BATCH_SIZE, shuffle=True)

定义网络

该网络包括两个卷积层和两个线性层,最后输出10个维度,即代表0-9十个数字。

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
class ConvNet(nn.Module):
  def __init__(self):
    super().__init__()
    self.conv1=nn.Conv2d(1,10,5) # input:(1,28,28) output:(10,24,24)
    self.conv2=nn.Conv2d(10,20,3) # input:(10,12,12) output:(20,10,10)
    self.fc1 = nn.Linear(20*10*10,500)
    self.fc2 = nn.Linear(500,10)
  def forward(self,x):
    in_size = x.size(0)
    out = self.conv1(x)
    out = F.relu(out)
    out = F.max_pool2d(out, 2, 2)
    out = self.conv2(out)
    out = F.relu(out)
    out = out.view(in_size,-1)
    out = self.fc1(out)
    out = F.relu(out)
    out = self.fc2(out)
    out = F.log_softmax(out,dim=1)
    return out

实例化网络

?
1
2
model = ConvNet().to(DEVICE) # 将网络移动到gpu上
optimizer = optim.Adam(model.parameters()) # 使用Adam优化器

定义训练函数

?
1
2
3
4
5
6
7
8
9
10
11
12
13
def train(model, device, train_loader, optimizer, 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+1)%30 == 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()))

定义测试函数

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
def test(model, device, test_loader):
  model.eval()
  test_loss = 0
  correct = 0
  with torch.no_grad():
    for data, target in test_loader:
      data, target = data.to(device), target.to(device)
      output = model(data)
      test_loss += F.nll_loss(output, target, reduction='sum').item() # 将一批的损失相加
      pred = output.max(1, keepdim=True)[1] # 找到概率最大的下标
      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)))

开始训练

?
1
2
3
for epoch in range(1, EPOCHS + 1):
  train(model, DEVICE, train_loader, optimizer, epoch)
  test(model, DEVICE, test_loader)

实验结果

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
Train Epoch: 1 [14848/60000 (25%)]  Loss: 0.375058
Train Epoch: 1 [30208/60000 (50%)]  Loss: 0.255248
Train Epoch: 1 [45568/60000 (75%)]  Loss: 0.128060
 
Test set: Average loss: 0.0992, Accuracy: 9690/10000 (97%)
 
Train Epoch: 2 [14848/60000 (25%)]  Loss: 0.093066
Train Epoch: 2 [30208/60000 (50%)]  Loss: 0.087888
Train Epoch: 2 [45568/60000 (75%)]  Loss: 0.068078
 
Test set: Average loss: 0.0599, Accuracy: 9816/10000 (98%)
 
Train Epoch: 3 [14848/60000 (25%)]  Loss: 0.043926
Train Epoch: 3 [30208/60000 (50%)]  Loss: 0.037321
Train Epoch: 3 [45568/60000 (75%)]  Loss: 0.068404
 
Test set: Average loss: 0.0416, Accuracy: 9859/10000 (99%)
 
Train Epoch: 4 [14848/60000 (25%)]  Loss: 0.031654
Train Epoch: 4 [30208/60000 (50%)]  Loss: 0.041341
Train Epoch: 4 [45568/60000 (75%)]  Loss: 0.036493
 
Test set: Average loss: 0.0361, Accuracy: 9873/10000 (99%)
 
Train Epoch: 5 [14848/60000 (25%)]  Loss: 0.027688
Train Epoch: 5 [30208/60000 (50%)]  Loss: 0.019488
Train Epoch: 5 [45568/60000 (75%)]  Loss: 0.018023
 
Test set: Average loss: 0.0344, Accuracy: 9875/10000 (99%)
 
Train Epoch: 6 [14848/60000 (25%)]  Loss: 0.024212
Train Epoch: 6 [30208/60000 (50%)]  Loss: 0.018689
Train Epoch: 6 [45568/60000 (75%)]  Loss: 0.040412
 
Test set: Average loss: 0.0350, Accuracy: 9879/10000 (99%)
 
Train Epoch: 7 [14848/60000 (25%)]  Loss: 0.030426
Train Epoch: 7 [30208/60000 (50%)]  Loss: 0.026939
Train Epoch: 7 [45568/60000 (75%)]  Loss: 0.010722
 
Test set: Average loss: 0.0287, Accuracy: 9892/10000 (99%)
 
Train Epoch: 8 [14848/60000 (25%)]  Loss: 0.021109
Train Epoch: 8 [30208/60000 (50%)]  Loss: 0.034845
Train Epoch: 8 [45568/60000 (75%)]  Loss: 0.011223
 
Test set: Average loss: 0.0299, Accuracy: 9904/10000 (99%)
 
Train Epoch: 9 [14848/60000 (25%)]  Loss: 0.011391
Train Epoch: 9 [30208/60000 (50%)]  Loss: 0.008091
Train Epoch: 9 [45568/60000 (75%)]  Loss: 0.039870
 
Test set: Average loss: 0.0341, Accuracy: 9890/10000 (99%)
 
Train Epoch: 10 [14848/60000 (25%)] Loss: 0.026813
Train Epoch: 10 [30208/60000 (50%)] Loss: 0.011159
Train Epoch: 10 [45568/60000 (75%)] Loss: 0.024884
 
Test set: Average loss: 0.0286, Accuracy: 9901/10000 (99%)
 
Train Epoch: 11 [14848/60000 (25%)] Loss: 0.006420
Train Epoch: 11 [30208/60000 (50%)] Loss: 0.003641
Train Epoch: 11 [45568/60000 (75%)] Loss: 0.003402
 
Test set: Average loss: 0.0377, Accuracy: 9894/10000 (99%)
 
Train Epoch: 12 [14848/60000 (25%)] Loss: 0.006866
Train Epoch: 12 [30208/60000 (50%)] Loss: 0.012617
Train Epoch: 12 [45568/60000 (75%)] Loss: 0.008548
 
Test set: Average loss: 0.0311, Accuracy: 9908/10000 (99%)
 
Train Epoch: 13 [14848/60000 (25%)] Loss: 0.010539
Train Epoch: 13 [30208/60000 (50%)] Loss: 0.002952
Train Epoch: 13 [45568/60000 (75%)] Loss: 0.002313
 
Test set: Average loss: 0.0293, Accuracy: 9905/10000 (99%)
 
Train Epoch: 14 [14848/60000 (25%)] Loss: 0.002100
Train Epoch: 14 [30208/60000 (50%)] Loss: 0.000779
Train Epoch: 14 [45568/60000 (75%)] Loss: 0.005952
 
Test set: Average loss: 0.0335, Accuracy: 9897/10000 (99%)
 
Train Epoch: 15 [14848/60000 (25%)] Loss: 0.006053
Train Epoch: 15 [30208/60000 (50%)] Loss: 0.002559
Train Epoch: 15 [45568/60000 (75%)] Loss: 0.002555
 
Test set: Average loss: 0.0357, Accuracy: 9894/10000 (99%)
 
Train Epoch: 16 [14848/60000 (25%)] Loss: 0.000895
Train Epoch: 16 [30208/60000 (50%)] Loss: 0.004923
Train Epoch: 16 [45568/60000 (75%)] Loss: 0.002339
 
Test set: Average loss: 0.0400, Accuracy: 9893/10000 (99%)
 
Train Epoch: 17 [14848/60000 (25%)] Loss: 0.004136
Train Epoch: 17 [30208/60000 (50%)] Loss: 0.000927
Train Epoch: 17 [45568/60000 (75%)] Loss: 0.002084
 
Test set: Average loss: 0.0353, Accuracy: 9895/10000 (99%)
 
Train Epoch: 18 [14848/60000 (25%)] Loss: 0.004508
Train Epoch: 18 [30208/60000 (50%)] Loss: 0.001272
Train Epoch: 18 [45568/60000 (75%)] Loss: 0.000543
 
Test set: Average loss: 0.0380, Accuracy: 9894/10000 (99%)
 
Train Epoch: 19 [14848/60000 (25%)] Loss: 0.001699
Train Epoch: 19 [30208/60000 (50%)] Loss: 0.000661
Train Epoch: 19 [45568/60000 (75%)] Loss: 0.000275
 
Test set: Average loss: 0.0339, Accuracy: 9905/10000 (99%)
 
Train Epoch: 20 [14848/60000 (25%)] Loss: 0.000441
Train Epoch: 20 [30208/60000 (50%)] Loss: 0.000695
Train Epoch: 20 [45568/60000 (75%)] Loss: 0.000467
 
Test set: Average loss: 0.0396, Accuracy: 9894/10000 (99%)

总结

一个实际项目的工作流程:找到数据集,对数据做预处理,定义我们的模型,调整超参数,测试训练,再通过训练结果对超参数进行调整或者对模型进行调整。

以上这篇使用PyTorch实现MNIST手写体识别代码就是小编分享给大家的全部内容了,希望能给大家一个参考,也希望大家多多支持服务器之家。

原文链接:https://blog.csdn.net/abcgkj/article/details/100884143

相关文章

热门资讯

2022年最旺的微信头像大全 微信头像2022年最新版图片
2022年最旺的微信头像大全 微信头像2022年最新版图片 2022-01-10
蜘蛛侠3英雄无归3正片免费播放 蜘蛛侠3在线观看免费高清完整
蜘蛛侠3英雄无归3正片免费播放 蜘蛛侠3在线观看免费高清完整 2021-08-24
背刺什么意思 网络词语背刺是什么梗
背刺什么意思 网络词语背刺是什么梗 2020-05-22
yue是什么意思 网络流行语yue了是什么梗
yue是什么意思 网络流行语yue了是什么梗 2020-10-11
暖暖日本高清免费中文 暖暖在线观看免费完整版韩国
暖暖日本高清免费中文 暖暖在线观看免费完整版韩国 2021-05-08
返回顶部