刘二大人pytorch学习笔记

发表于 | 分类于 | 阅读次数：

pytorch学习记录

1. 加载文件

1
2
3
4
5
6
7
8
9
10
11
12
13
14
1.加载txt
file = np.loadtxt(filepath,delimiter=',',dtype=np.float32)
x_data = torch.from_numpy(file[:,:-1]) #不包含最后一列
y_data = torch.from_numpy(file[:,[-1]])  #只包含最后一列

2.加载h5文件
import h5py
file = h5py.File('.h5','r')
file.keys() 	#查看h5文件的keys
x=np.array(file['train_x'])
x = x.astype('float32')/255
x = torch.from_numpy(x)
# x的维度为 m * h * w * c,需要转换维度
x.permute(0,3,1,2)

2. 手写数字识别model搭建

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
import torch
import torchvision
from torch import nn
from torch.utils.data import DataLoader

train_data = torchvision.datasets.MNIST('data/mnist',transform=torchvision.transforms.ToTensor(),train=True,download=True)
test_data = torchvision.datasets.MNIST('data/mnist',transform=torchvision.transforms.ToTensor(),train=False,download=True)

train_loader = DataLoader(train_data,batch_size=64,shuffle=True)
test_loader = DataLoader(test_data,batch_size=64,shuffle=True)

train_len =  len(train_data)
test_len = len(test_data)

class Model(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(1,6,5)
        self.avg_pool1 = nn.AvgPool2d(2,stride=2)
        self.conv2 = nn.Conv2d(6,16,5)
        self.avg_pool2 = nn.AvgPool2d(2,stride=2)
        self.flatten = nn.Flatten()
        self.l1 = nn.Linear(256,64)
        self.l2 = nn.Linear(64,10)
        self.relu = nn.ReLU()

    def forward(self,x):
        x = self.conv1(x)
        x = self.avg_pool1(x)
        x = self.conv2(x)
        x = self.avg_pool2(x)
        x = self.flatten(x)
        x = self.relu(self.l1(x))
        x = self.l2(x)
        return x

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model = Model()
model= model.to(device)

loss_func = nn.CrossEntropyLoss()
loss_func = loss_func.to(device)
optim = torch.optim.SGD(model.parameters(),lr=0.01)

epoch = 10
for i in range(epoch):

    print('----------------第{}次迭代开始-----------'.format(i+1))
    loss_cnt = 0
    acc = 0
    for data in train_loader:
        imgs, labels = data      #64 1 28 28
       # imgs = torch.reshape(imgs, (-1, 784))
        imgs = imgs.to(device)
        labels = labels.to(device)

        y_hat = model(imgs)

        acc += (y_hat.argmax(1) == labels).sum()
        loss = loss_func(y_hat, labels)
        loss_cnt += loss

        optim.zero_grad()
        loss.backward()
        optim.step()

    print("第{}次的loss为{}".format(i + 1, loss_cnt))
    print("第{}次的acc为{}".format(i + 1, acc/train_len))

#test
with torch.no_grad():
    acc = 0
    for data in test_loader:
        imgs, labels = data
        #imgs: 64*1*32*32
       # imgs = torch.reshape(imgs, (-1,784))
        imgs = imgs.to(device)
        labels = labels.to(device)

        y_hat = model(imgs)

        acc += (y_hat.argmax(1) == labels).sum()

    print("acc={}".format(acc/test_len))