Pytorch 学习笔记 - Linear Regression (2)

Posted on | In , , |

读 «动⼿学深度学习 Pytorch 版»

1
2
3
4
5
6
7
8
import torch
import torch.nn as nn
import matplotlib.pyplot as plt
import torch.utils.data as Data

from IPython import display

%matplotlib inline

⽣成数据集

1
2
3
4
5
torch.manual_seed(1)
num_samples, num_features, num_outputs = 1000, 2, 1
true_w = torch.rand(num_features, 1)
true_b = torch.rand(1)
true_w, true_b # true w, b

1
2
3
(tensor([[0.7576],
         [0.2793]]),
 tensor([0.4031]))

1
2
X = torch.randn(num_samples, num_features) 
y = torch.mm(X, true_w) + true_b + torch.randn(num_samples,1) * 0.003

读取数据

1
2
dataset = Data.TensorDataset(X, y)
data_iter = Data.DataLoader(dataset, batch_size=10, shuffle=True)

定义模型

1
2
3
4
5
6
7
8
9
10
class LinearNet(nn.Module):
    def __init__(self, num_features):
        super(LinearNet, self).__init__()
        self.linear = nn.Linear(num_features, 1) # 1 output
        
    def forward(self, x):
        x = self.linear(x)
        return x
net = LinearNet(num_features)
print(net, '\n', net.linear, '\n', net.linear.weight, '\n', net.linear.bias)

1
2
3
4
5
6
7
8
LinearNet(
  (linear): Linear(in_features=2, out_features=1, bias=True)
) 
 Linear(in_features=2, out_features=1, bias=True) 
 Parameter containing:
tensor([[-0.0979, -0.0225]], requires_grad=True) 
 Parameter containing:
tensor([-0.1533], requires_grad=True)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
# # 写法一
# net = nn.Sequential(
#     nn.Linear(num_inputs, 1)
#     # 此处还可以传入其他层
#     )

# # 写法二
# net = nn.Sequential()
# net.add_module('linear', nn.Linear(num_inputs, 1))
# # net.add_module ......

# # 写法三
# from collections import OrderedDict
# net = nn.Sequential(OrderedDict([
#           ('linear', nn.Linear(num_inputs, 1))
#           # ......
#         ]))

# print(net)
# print(net[0])

1
2
for param in net.parameters():
    print(param)

1
2
3
4
Parameter containing:
tensor([[-0.0979, -0.0225]], requires_grad=True)
Parameter containing:
tensor([-0.1533], requires_grad=True)

注意: torch.nn 仅⽀持输⼊一个batch的样本不⽀持单个样本输入，如果只有单个样本，可使用 input.unsqueeze(0) 来添加一维。

初始化模型参数

1
2
3
4
5
6
from torch.nn import init

init.normal_(net.linear.weight, mean=0, std=0.01)
init.constant_(net.linear.bias, val=0)  
# 也可以直接修改bias的data: net[0].bias.data.fill_(0)
print(net.linear.weight, net.linear.bias)

1
2
3
Parameter containing:
tensor([[ 0.0042, -0.0070]], requires_grad=True) Parameter containing:
tensor([0.], requires_grad=True)

定义损失函数

1
loss = nn.MSELoss()

定义优化算法

1
2
3
4
5
import torch.optim as optim

optimizer = optim.SGD(net.parameters(), lr=1e-3)
print(optimizer)


1
2
3
4
5
6
7
8
SGD (
Parameter Group 0
    dampening: 0
    lr: 0.001
    momentum: 0
    nesterov: False
    weight_decay: 0
)

我们还可以为不同子网络设置不同的学习率，这在finetune时经常用到。例：

1
2
3
4
5
6
# optimizer =optim.SGD([
#                 # 如果对某个参数不指定学习率，就使用最外层的默认学习率
#                 {'params': net.subnet1.parameters()}, # lr=0.03
#                 {'params': net.subnet2.parameters(), 'lr': 0.01}
#             ], lr=0.03)

训练模型

1
2
3
4
5
6
7
8
9
10
11
12
13
num_epochs = 100
train_loss = []
for epoch in range(num_epochs):
    for bx, by in data_iter:
        output = net(bx)
        l = loss(output, by)
        optimizer.zero_grad() # 梯度清零，等价于net.zero_grad()
        l.backward()
        optimizer.step()
    with torch.no_grad():
        train_loss.append(loss(net(X), y).item())
    if (epoch + 1) % 10 == 0:
        print('epoch %d, loss: %f' % (epoch, train_loss[-1]))

1
2
3
4
5
6
7
8
9
10
epoch 9, loss: 0.012866
epoch 19, loss: 0.000209
epoch 29, loss: 0.000012
epoch 39, loss: 0.000009
epoch 49, loss: 0.000009
epoch 59, loss: 0.000009
epoch 69, loss: 0.000009
epoch 79, loss: 0.000009
epoch 89, loss: 0.000009
epoch 99, loss: 0.000009

1
2
3
plt.figure(figsize=(8,6))
plt.plot(train_loss)
plt.show()

cost

1
print(net.linear.weight, net.linear.bias)

1
2
3
Parameter containing:
tensor([[0.7577, 0.2792]], requires_grad=True) Parameter containing:
tensor([0.4031], requires_grad=True)

3D数据图

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
import plotly.graph_objects as go

input_data = go.Scatter3d(
    x = X[:,0].numpy(),
    y = X[:,1].numpy(),
    z = y[:,0].numpy(),
    mode = 'markers',
    name = 'Input Data',
    marker = dict(
        color = 'rgb(160, 57, 147)',
        size = 4,
        symbol = 'circle',
        opacity = 0.9
    )
)

with torch.no_grad():
    y_hat = net(X)
    
predict_data = go.Scatter3d(
    x = X[:,0].numpy(),
    y = X[:,1].numpy(),
    z = y_hat.detach().numpy()[:,0],
    mode = 'markers',
    name = 'Predict Data',
    marker = dict(
        color = 'rgb(50, 168, 139)',
        size = 2,
        symbol = 'square',
        opacity = 0.9
    )
)

data = [input_data, predict_data]
layout = go.Layout(
    scene = dict(
        xaxis = dict(title='X1'),
        yaxis = dict(title='X2'),
        zaxis = dict(title='y'),
    ),
    margin = dict(
        l=10,
        r=10,
        b=10,
        t=10
    ),
    width = 620,
    height = 440,
)

fig = go.Figure(data = data, layout = layout)
fig.show()