pytorch实现线性回归_Python

pytorch实现线性回归

2021-10-08 00:16逝去〃年华 Python

这篇文章主要为大家详细介绍了pytorch实现线性回归，文中示例代码介绍的非常详细，具有一定的参考价值，感兴趣的小伙伴们可以参考一下

pytorch实现线性回归代码练习实例，供大家参考，具体内容如下

欢迎大家指正，希望可以通过小的练习提升对于pytorch的掌握

				?

									# 随机初始化一个二维数据集，使用朋友torch训练一个回归模型

									import numpy as np

									import random

									import matplotlib.pyplot as plt

									x = np.arange(20)

									y = np.array([5*x[i] + random.randint(1,20) for i in range(len(x))])    # random.randint(参数1,参数2)函数返回参数1和参数2之间的任意整数

									print('-'*50)

									# 打印数据集

									print(x)

									print(y)

									import torch

									x_train = torch.from_numpy(x).float()

									y_train = torch.from_numpy(y).float()

									# model

									class linearregression(torch.nn.module):

									    def __init__(self):

									        super(linearregression, self).__init__()

									        # 输入与输出都是一维的

									        self.linear = torch.nn.linear(1,1)

									    def forward(self,x):

									        return self.linear(x)

									# 新建模型，误差函数，优化器

									model = linearregression()

									criterion = torch.nn.mseloss()

									optimizer = torch.optim.sgd(model.parameters(),0.001)

									# 开始训练

									num_epoch = 20

									for i in range(num_epoch):

									    input_data = x_train.unsqueeze(1)

									    target = y_train.unsqueeze(1)           # unsqueeze(1)在第二维增加一个维度

									    out = model(input_data)

									    loss = criterion(out,target)

									    optimizer.zero_grad()

									    loss.backward()

									    optimizer.step()

									    print("eopch:[{}/{},loss:[{:.4f}]".format(i+1,num_epoch,loss.item()))

									    if ((i+1)%2 == 0):

									        predict = model(input_data)

									        plt.plot(x_train.data.numpy(),predict.squeeze(1).data.numpy(),"r")

									        loss = criterion(predict,target)

									        plt.title("loss:{:.4f}".format(loss.item()))

									        plt.xlabel("x")

									        plt.ylabel("y")

									        plt.scatter(x_train,y_train)

									        plt.show()