b站刘二视频,地址:
https://www.bilibili.com/video/BV1Y7411d7Ys?p=9&vd_source=79d752a233297190ff0b01ca81ccd878
代码(课中作业)
还是上节课的糖尿病的二元分类问题,四步法构造
import torch
import numpy as np
import matplotlib.pyplot as plt
from torch.utils.data import Dataset
from torch.utils.data import DataLoader
#-------------------------------------------step1 prepare data----------------------------------------
class Data(Dataset): #构造自己的类,继承自Dataset类
def __init__(self, filepath):
xy = np.loadtxt(filepath, delimiter=',', dtype=np.float32)
#文件路径 数据中的分隔符 数据类型
self.len = xy.shape[0] #返回几行几列(矩阵)
self.x_data = torch.from_numpy(xy[:, :-1])#取每一行,取到最后一列
self.y_data = torch.from_numpy(xy[:, [-1]])#取每一行,取最后一列
def __getitem__(self, item): #获取某一行元素
return self.x_data[item], self.y_data[item]
def __len__(self):
return self.len
dataset = Data('diabetes.csv')
dataloader = DataLoader(dataset=dataset, batch_size=32, shuffle=True)
#------------------------------------------setp2 design model--------------------------------------------
class Modle(torch.nn.Module):
def __init__(self):
super(Modle, self).__init__()
self.linear1 = torch.nn.Linear(8, 6)
self.linear2 = torch.nn.Linear(6, 4)
self.linear3 = torch.nn.Linear(4, 1)
self.sigmoid = torch.nn.Sigmoid()
def forward(self, x):
x = self.sigmoid(self.linear1(x))
x = self.sigmoid(self.linear2(x))
x = self.sigmoid(self.linear3(x))
return x
model = Modle()
#---------------------------------------step3 constuct loss and optimizer-----------------------------
criteration = torch.nn.BCELoss(reduction='mean')#mean求均值
optimizer = torch.optim.SGD(model.parameters(), lr=0.0001) #要更新的是model中的参数, 学习率
#---------------------------------------step4 traning cycle------------------------------------------
if __name__ == '__main__': #要写这个
loss_lst = []
for epoch in range(1000):#外层epoch是所有数据集都遍历过的一次训练
sum = 0#
for i, data in enumerate(dataloader, 0):#一次是一个batch,从整个数据集中的一部分
inputs, lables = data#x y
y_pred = model(inputs)
loss = criteration(y_pred, lables)
sum += loss.item()
optimizer.zero_grad()
loss.backward()
optimizer.step()
loss_lst.append(sum / dataloader.batch_size)
#可视化
num_lst = [i for i in range(len(loss_lst))]
plt.plot(num_lst, loss_lst)
plt.xlabel("epoch")
plt.ylabel("loss")
plt.show()
运行结果
MINIST数据导入