PyTorch笔记 入门：写一个简单的神经网络2：Classification（分类）

相关视频：
PyTorch 动态神经网络 (莫烦 Python 教学)

一、导入库、创造数据

二、编写网络

1.设计网络、编写代码、创建网络

写法一

同上一篇博客PyTorch笔记 入门：写一个简单的神经网络1：Regression（回归）相同

写法二

用torch.nn.Sequential：

2.优化器

3.训练网络

4.画出loss的图像

5.预测

三、完整代码

写法一

import torch
import torch.nn.functional as F
from torch.autograd import Variable
import matplotlib.pyplot as plt

# 自己创建的网路
class Net(torch.nn.Module):
    def __init__(self, n_input, n_hidden, n_output):
        super(Net, self).__init__()
        self.hidden_layer = torch.nn.Linear(n_input, n_hidden)
        self.output_layer = torch.nn.Linear(n_hidden, n_output)
    
    def forward(self, x):
        x = F.relu(self.hidden_layer(x))
        x = self.output_layer(x)
        return x

n_data = torch.ones(100, 2)

# 第0类
x0 = torch.normal(2*n_data, 1)
y0 = torch.zeros(100) # 标签

# 第1类
x1 = torch.normal(-2*n_data, 1)
y1 = torch.ones(100) # 标签

x = torch.cat((x0, x1), 0).type(torch.FloatTensor)
y = torch.cat((y0, y1)).type(torch.LongTensor)

x, y = Variable(x), Variable(y)

print(x.shape)
print(y.shape)

plt.scatter(x.data.numpy()[0:100,0], x.data.numpy()[0:100,1], c='r') # 第0类
plt.scatter(x.data.numpy()[100:200,0], x.data.numpy()[100:200,1], c='b') # 第1类
plt.show()

# 创建网络
net = Net(2, 10, 2)
print(net)

# 优化器
optimizer = torch.optim.SGD(net.parameters(), lr=0.5)

# 损失函数
loss_func = torch.nn.CrossEntropyLoss() # Cross Entropy Loss

# 训练
losses = []
for i in range(100):
    # 计算loss
    output = net(x)
    loss = loss_func(output, y)
    # 训练优化
    optimizer.zero_grad()
    loss.backward()
    optimizer.step()
    # 每训练5次输出一次loss
    if i%5 == 0:
        print(loss.data.item())
        losses.append(loss.data.item())
        
# 画出loss的图像
import matplotlib.pyplot as plt

plt.plot(range(len(losses)), losses)
plt.xlabel('epoch')
plt.ylabel('loss')
plt.xticks(range(0,20,2))
plt.show()

# 预测
test = torch.randn(30, 2)
print(test.shape)

out = net(test)
prediction = torch.max(F.softmax(out), 1)[1]
result = prediction.data.numpy()
print(result.shape)

colors = ['r', 'b']
for i in range(test.shape[0]):
    plt.scatter(test[i,0], test[i,1], c=colors[result[i]])
plt.show()

写法二

import torch
import torch.nn.functional as F
from torch.autograd import Variable
import matplotlib.pyplot as plt

n_data = torch.ones(100, 2)

# 第0类
x0 = torch.normal(2*n_data, 1)
y0 = torch.zeros(100) # 标签

# 第1类
x1 = torch.normal(-2*n_data, 1)
y1 = torch.ones(100) # 标签

x = torch.cat((x0, x1), 0).type(torch.FloatTensor)
y = torch.cat((y0, y1)).type(torch.LongTensor)

x, y = Variable(x), Variable(y)

print(x.shape)
print(y.shape)

plt.scatter(x.data.numpy()[0:100,0], x.data.numpy()[0:100,1], c='r') # 第0类
plt.scatter(x.data.numpy()[100:200,0], x.data.numpy()[100:200,1], c='b') # 第1类
plt.show()

# 创建网络
net = torch.nn.Sequential(
    torch.nn.Linear(2, 10),
    torch.nn.ReLU(),
    torch.nn.Linear(10,2),
)
print(net)

# 优化器
optimizer = torch.optim.SGD(net.parameters(), lr=0.5)

# 损失函数
loss_func = torch.nn.CrossEntropyLoss() # Cross Entropy Loss

# 训练
losses = []
for i in range(100):
    # 计算loss
    output = net(x)
    loss = loss_func(output, y)
    # 训练优化
    optimizer.zero_grad()
    loss.backward()
    optimizer.step()
    # 每训练5次输出一次loss
    if i%5 == 0:
        print(loss.data.item())
        losses.append(loss.data.item())
        
# 画出loss的图像
import matplotlib.pyplot as plt

plt.plot(range(len(losses)), losses)
plt.xlabel('epoch')
plt.ylabel('loss')
plt.xticks(range(0,20,2))
plt.show()

# 预测
test = torch.randn(30, 2)
print(test.shape)

out = net(test)
prediction = torch.max(F.softmax(out), 1)[1]
result = prediction.data.numpy()
print(result.shape)

colors = ['r', 'b']
for i in range(test.shape[0]):
    plt.scatter(test[i,0], test[i,1], c=colors[result[i]])
plt.show()