import torch.nn as nn    # 神經(jīng)網(wǎng)絡(luò)庫
import matplotlib.pyplot as plt  # 繪圖
import torch        # 張量
from torch import optim  # 優(yōu)化器庫
import numpy as np          # 數(shù)據(jù)處理
xy = np.loadtxt('./diabetes.csv.gz',delimiter=',',dtype=np.float32)    # 加載數(shù)據(jù)集
x_data = torch.from_numpy(xy[:,:-1])  # 所有行，除了最后一列的元素
y_data = torch.from_numpy(xy[:,-1]).view(-1,1) # -1也能拿出來是向量，但是[-1]會保證拿出來的是個矩陣
epoch_list =[]
loss_list = []
class Model(nn.Module):
    def __init__(self):
        super(Model,self).__init__()
        self.linear1 = nn.Linear(8,6)
        self.linear2 = nn.Linear(6,3)
        self.linear3 = nn.Linear(3,1)
        self.sigmoid = nn.Sigmoid()
        self.relu = nn.ReLU()
    def forward(self,x):
        x = self.relu(self.linear1(x))
        x = self.relu(self.linear2(x))
        x = self.sigmoid(self.linear3(x))
        return x
model = Model()
criterion = nn.BCELoss()
optimizer = optim.SGD(model.parameters(),lr =0.01)
for epoch in range(1000):
    y_pred = model(x_data)
    loss = criterion(y_pred,y_data)   # 計算損失
    if epoch % 100 ==0:   # 每隔100次打印一下
        print(epoch,loss.item())
    #back propagation
    optimizer.zero_grad()    # 梯度清零
    loss.backward()          # 反向傳播
    optimizer.step()         # 梯度更新
    epoch_list.append(epoch)
    loss_list.append(loss.item())
plt.plot(epoch_list,loss_list)
plt.show()