import torch
import numpy as np

## 載入數據集，delimiter--分隔符
xy = np.loadtxt('diabetes.csv.gz', delimiter=',', dtype=np.float32)
#從numpy中生成Tensor
x_data = torch.from_numpy(xy[:, :-1])
y_data = torch.from_numpy(xy[:, [-1]])



##Design Model

##構造類，繼承torch.nn.Module類
class Model(torch.nn.Module):
    ## 構造函數，初始化對象
    def __init__(self):
        ##super調用父類
        super(Model, self).__init__()
        ##構造三層神經網絡
        self.linear1 = torch.nn.Linear(8, 6)
        self.linear2 = torch.nn.Linear(6, 4)
        self.linear3 = torch.nn.Linear(4, 1)
        ##激活函數，進行非線性變換
        self.sigmoid = torch.nn.Sigmoid()
        
    ## 構造函數，前饋運算
    def forward(self, x):
        x = self.sigmoid(self.linear1(x))
        x = self.sigmoid(self.linear2(x))
        x = self.sigmoid(self.linear3(x))
        return x
    
# =============================================================================
#     # 激活函數，進行非線性變換
#         self.activate = torch.nn.ReLU()
#         
#     # 構造函數，前饋運算
#     def forward(self, x):
#         x = self.activate(self.linear1(x))
#         x = self.activate(self.linear2(x))
#         #最后一層為了保證輸出結果（概率）在[0,1]，要用sigmoid
#         x = self.sigmoid(self.linear3(x))
#         return x
# =============================================================================
    
model = Model()

##Construct Loss and Optimizer

##損失函數，傳入y和y_pred，size_average--是否取平均
criterion = torch.nn.BCELoss(size_average = True)

##優化器，model.parameters()找出模型所有的參數，Lr--學習率
optimizer = torch.optim.SGD(model.parameters(), lr=0.1)


## Training cycle

for epoch in range(100):
    ##前向傳播
    y_pred = model(x_data)
    loss = criterion(y_pred, y_data)
    print(epoch, loss.item())
    
    ##梯度歸零
    optimizer.zero_grad()
    ##反向傳播
    loss.backward()
    ##更新
    optimizer.step()