本帖最后由 梦缠绕的时候 于 2019-6-19 14:42 编辑
1.引文
深度学习的比赛中，图片分类是很常见的比赛，同时也是很难取得特别高名次的比赛，因为图片分类已经被大家研究的很透彻，一些开源的网络很容易取得高分。如果大家还掌握不了使用开源的网络进行训练，再慢慢去模型调优，很难取得较好的成绩。
我们在[PyTorch小试牛刀]实战六·准备自己的数据集用于训练讲解了如何制作自己的数据集用于训练，这个教程在此基础上，进行训练与应用。
2.数据介绍
数据下载地址： https://download.csdn.net/download/xiaosongshine/11128410
这次的实战使用的数据是交通标志数据集，共有62类交通标志。其中训练集数据有4572张照片（每个类别大概七十个），测试数据集有2520张照片（每个类别大概40个）。数据包含两个子目录分别train与test：
为什么还需要测试数据集呢？这个测试数据集不会拿来训练，是用来进行模型的评估与调优。
train与test每个文件夹里又有62个子文件夹，每个类别在同一个文件夹内：
我从中打开一个文件间，把里面图片展示出来：
其中每张照片都类似下面的例子，100*100*3的大小。100是照片的照片的长和宽，3是什么呢？这其实是照片的色彩通道数目，RGB。彩色照片存储在计算机里就是以三维数组的形式。我们送入网络的也是这些数组。
3.网络构建
1.导入Python包，定义一些参数
1import torch as t
2import torchvision as tv
3import os
4import time
5import numpy as np
6from tqdm import tqdm
7
8
9class DefaultConfigs(object):
10
11 data_dir = "./traffic-sign/"
12 data_list = ["train","test"]
13
14 lr = 0.001
15 epochs = 10
16 num_classes = 62
17 image_size = 224
18 batch_size = 40
19 channels = 3
20 gpu = "0"
21 train_len = 4572
22 test_len = 2520
23 use_gpu = t.cuda.is_available()
24
25config = DefaultConfigs()
2.数据准备，采用PyTorch提供的读取方式
注意一点Train数据需要进行随机裁剪，Test数据不要进行裁剪了
1normalize = tv.transforms.Normalize(mean = [0.485, 0.456, 0.406],
2 std = [0.229, 0.224, 0.225]
3 )
4
5transform = {
6 config.data_list[0]:tv.transforms.Compose(
7 [tv.transforms.Resize([224,224]),tv.transforms.CenterCrop([224,224]),
8 tv.transforms.ToTensor(),normalize]#tv.transforms.Resize 用于重设图片大小
9 ) ,
10 config.data_list[1]:tv.transforms.Compose(
11 [tv.transforms.Resize([224,224]),tv.transforms.ToTensor(),normalize]
12 )
13}
14
15datasets = {
16 x:tv.datasets.ImageFolder(root = os.path.join(config.data_dir,x),transform=transform[x])
17 for x in config.data_list
18}
19
20dataloader = {
21 x:t.utils.data.DataLoader(dataset= datasets[x],
22 batch_size=config.batch_size,
23 shuffle=True
24 )
25 for x in config.data_list
26}
3.构建网络模型（使用resnet18进行迁移学习，训练参数为最后一个全连接层 t.nn.Linear(512,num_classes)）
1def get_model(num_classes):
2
3 model = tv.models.resnet18(pretrained=True)
4 for parma in model.parameters():
5 parma.requires_grad = False
6 model.fc = t.nn.Sequential(
7 t.nn.Dropout(p=0.3),
8 t.nn.Linear(512,num_classes)
9 )
10 return(model)
如果电脑硬件支持，可以把下述代码屏蔽，则训练整个网络，最终准确率会上升，训练数据会变慢。
1for parma in model.parameters():
2 parma.requires_grad = False
模型输出
1ResNet(
2 (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
3 (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
4 (relu): ReLU(inplace)
5 (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
6 (layer1): Sequential(
7 (0): BasicBlock(
8 (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
9 (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
10 (relu): ReLU(inplace)
11 (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
12 (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
13 )
14 (1): BasicBlock(
15 (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
16 (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
17 (relu): ReLU(inplace)
18 (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
19 (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
20 )
21 )
22 (layer2): Sequential(
23 (0): BasicBlock(
24 (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
25 (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
26 (relu): ReLU(inplace)
27 (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
28 (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
29 (downsample): Sequential(
30 (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)
31 (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
32 )
33 )
34 (1): BasicBlock(
35 (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
36 (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
37 (relu): ReLU(inplace)
38 (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
39 (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
40 )
41 )
42 (layer3): Sequential(
43 (0): BasicBlock(
44 (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
45 (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
46 (relu): ReLU(inplace)
47 (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
48 (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
49 (downsample): Sequential(
50 (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)
51 (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
52 )
53 )
54 (1): BasicBlock(
55 (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
56 (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
57 (relu): ReLU(inplace)
58 (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
59 (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
60 )
61 )
62 (layer4): Sequential(
63 (0): BasicBlock(
64 (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
65 (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
66 (relu): ReLU(inplace)
67 (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
68 (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
69 (downsample): Sequential(
70 (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
71 (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
72 )
73 )
74 (1): BasicBlock(
75 (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
76 (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
77 (relu): ReLU(inplace)
78 (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
79 (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
80 )
81 )
82 (avgpool): AvgPool2d(kernel_size=7, stride=1, padding=0)
83 (fc): Sequential(
84 (0): Dropout(p=0.3)
85 (1): Linear(in_features=512, out_features=62, bias=True)
86 )
87)
4.训练模型（支持自动GPU加速）
1def train(epochs):
2
3 model = get_model(config.num_classes)
4 print(model)
5 loss_f = t.nn.CrossEntropyLoss()
6 if(config.use_gpu):
7 model = model.cuda()
8 loss_f = loss_f.cuda()
9
10 opt = t.optim.Adam(model.fc.parameters(),lr = config.lr)
11 time_start = time.time()
12
13 for epoch in range(epochs):
14 train_loss = []
15 train_acc = []
16 test_loss = []
17 test_acc = []
18 model.train(True)
19 print("Epoch {}/{}".format(epoch+1,epochs))
20 for batch, datas in tqdm(enumerate(iter(dataloader["train"]))):
21 x,y = datas
22 if (config.use_gpu):
23 x,y = x.cuda(),y.cuda()
24 y_ = model(x)
25 #print(x.shape,y.shape,y_.shape)
26 _, pre_y_ = t.max(y_,1)
27 pre_y = y
28 #print(y_.shape)
29 loss = loss_f(y_,pre_y)
30 #print(y_.shape)
31 acc = t.sum(pre_y_ == pre_y)
32
33 loss.backward()
34 opt.step()
35 opt.zero_grad()
36 if(config.use_gpu):
37 loss = loss.cpu()
38 acc = acc.cpu()
39 train_loss.append(loss.data)
40 train_acc.append(acc)
41 #if((batch+1)%5 ==0):
42 time_end = time.time()
43 print("Batch {}, Train loss:{:.4f}, Train acc:{:.4f}, Time: {}"\
44 .format(batch+1,np.mean(train_loss)/config.batch_size,np.mean(train_acc)/config.batch_size,(time_end-time_start)))
45 time_start = time.time()
46
47 model.train(False)
48 for batch, datas in tqdm(enumerate(iter(dataloader["test"]))):
49 x,y = datas
50 if (config.use_gpu):
51 x,y = x.cuda(),y.cuda()
52 y_ = model(x)
53 #print(x.shape,y.shape,y_.shape)
54 _, pre_y_ = t.max(y_,1)
55 pre_y = y
56 #print(y_.shape)
57 loss = loss_f(y_,pre_y)
58 acc = t.sum(pre_y_ == pre_y)
59
60 if(config.use_gpu):
61 loss = loss.cpu()
62 acc = acc.cpu()
63
64 test_loss.append(loss.data)
65 test_acc.append(acc)
66 print("Batch {}, Test loss:{:.4f}, Test acc:{:.4f}".format(batch+1,np.mean(test_loss)/config.batch_size,np.mean(test_acc)/config.batch_size))
67
68 t.save(model,str(epoch+1)+"ttmodel.pkl")
69
70
71
72if __name__ == "__main__":
73 train(config.epochs)
训练结果如下：
1Epoch 1/10
2115it [00:48, 2.63it/s]
3Batch 115, Train loss:0.0590, Train acc:0.4635, Time: 48.985504150390625
463it [00:24, 2.62it/s]
5Batch 63, Test loss:0.0374, Test acc:0.6790, Time :24.648272275924683
6Epoch 2/10
7115it [00:45, 3.22it/s]
8Batch 115, Train loss:0.0271, Train acc:0.7576, Time: 45.68823838233948
963it [00:23, 2.62it/s]
10Batch 63, Test loss:0.0255, Test acc:0.7524, Time :23.271782875061035
11Epoch 3/10
12115it [00:45, 3.19it/s]
13Batch 115, Train loss:0.0181, Train acc:0.8300, Time: 45.92648506164551
1463it [00:23, 2.60it/s]
15Batch 63, Test loss:0.0212, Test acc:0.7861, Time :23.80789279937744
16Epoch 4/10
17115it [00:45, 3.28it/s]
18Batch 115, Train loss:0.0138, Train acc:0.8767, Time: 45.27525019645691
1963it [00:23, 2.57it/s]
20Batch 63, Test loss:0.0173, Test acc:0.8385, Time :23.736321449279785
21Epoch 5/10
22115it [00:44, 3.22it/s]
23Batch 115, Train loss:0.0112, Train acc:0.8950, Time: 44.983638286590576
2463it [00:22, 2.69it/s]
25Batch 63, Test loss:0.0156, Test acc:0.8520, Time :22.790074348449707
26Epoch 6/10
27115it [00:44, 3.19it/s]
28Batch 115, Train loss:0.0095, Train acc:0.9159, Time: 45.10426950454712
2963it [00:22, 2.77it/s]
30Batch 63, Test loss:0.0158, Test acc:0.8214, Time :22.80412459373474
31Epoch 7/10
32115it [00:45, 2.95it/s]
33Batch 115, Train loss:0.0081, Train acc:0.9280, Time: 45.30439043045044
3463it [00:23, 2.66it/s]
35Batch 63, Test loss:0.0139, Test acc:0.8528, Time :23.122379541397095
36Epoch 8/10
37115it [00:44, 3.23it/s]
38Batch 115, Train loss:0.0073, Train acc:0.9300, Time: 44.304762840270996
3963it [00:22, 2.74it/s]
40Batch 63, Test loss:0.0142, Test acc:0.8496, Time :22.801835536956787
41Epoch 9/10
42115it [00:43, 3.19it/s]
43Batch 115, Train loss:0.0068, Train acc:0.9361, Time: 44.08414030075073
4463it [00:23, 2.44it/s]
45Batch 63, Test loss:0.0142, Test acc:0.8437, Time :23.604419231414795
46Epoch 10/10
47115it [00:46, 3.12it/s]
48Batch 115, Train loss:0.0063, Train acc:0.9337, Time: 46.76597046852112
4963it [00:24, 2.65it/s]
50Batch 63, Test loss:0.0130, Test acc:0.8591, Time :24.64351773262024
训练10个Epoch，测试集准确率可以到达0.86，已经达到不错效果。通过修改参数，增加训练，可以达到更高的准确率。
|
收藏
淘帖
踩
感谢分享~~
