pytorch实现线性回归
前言
线性回归是监督学习里面一个非常简单的模型,同时梯度下降也是深度学习中应用最广的优化算法
梯度下降法
实在是一些公式以文本格式打不出来,所以采用截图!!!!
一元线性回归
首先构造初始数据(小编感觉这时候画图更清晰)
import numpy as np
import matplotlib.pyplot as plt
if __name__ == "__main__":
# 设置数据 x 和 y
x_train = np.array([[3.3], [4.4], [5.5], [6.71], [6.93], [4.168],
[9.779], [6.182], [7.59], [2.167], [7.042],
[10.791], [5.313], [7.997], [3.1]], dtype=np.float32)
y_train = np.array([[1.7], [2.76], [2.09], [3.19], [1.694], [1.573],
[3.366], [2.596], [2.53], [1.221], [2.827],
[3.465], [1.65], [2.904], [1.3]], dtype=np.float32)
plt.scatter(x_train,y_train)
plt.show()
画出没有更新之前的预测模型和真实的模型之间的对比
import torch
from torch.autograd import Variable
import numpy as np
import matplotlib.pyplot as plt
class LinearModel(object):
def __init__(self,x_train,y_train):
self.x_train = x_train
self.y_train = y_train
def get_loss(self,y_predict,y_train):
return torch.mean((y_predict - y_train) ** 2 )
def single(self):
'''
单一变量的线性回归
:return: None
'''
# 将数据从ndarray类型转化为tensor
x_train = torch.from_numpy(self.x_train)
y_train = torch.from_numpy(self.y_train)
# 定义权重w 和 偏置b
# 权重随机生成
w = Variable(torch.rand(1), requires_grad=True)
# 偏置初始化为0
b = Variable(torch.zeros(1), requires_grad=True)
# 开始建立线性回归模型
# y = w * x+b
x_train = Variable(x_train)
y_train = Variable(y_train)
y_predict = x_train*w + b
# # 绘制图像,看下模型的输出结果
# plt.scatter(x_train.data.numpy(), y_train.data.numpy(), label="real")
# plt.scatter(x_train.data.numpy(), y_predict.data.numpy(), label="predict")
# plt.legend
# plt.show()
loss = self.get_loss(y_predict=y_predict,y_train=y_train)
print(loss)
# 求导
loss.backward()
# 查看w 和b 的梯度
print(w.grad)
print(b.grad)
# 更新一次参数
w.data = w.data - 1e-2 * w.grad.data
b.data = b.data - 1e-2 * b.grad.data
# 绘制图像,看下模型的输出结果
plt.scatter(x_train.data.numpy(), y_train.data.numpy(), label="real")
plt.scatter(x_train.data