Question

它运行良好，但没有学到任何东西。损耗永远不会减少。此外，它还会发生运行时警告：功率除以零。


phi = np.array([0, 0.00724873038246, 0.0126062332691, 0.0179158239032, 0.0237780447692, 0.0301819694947, 0.037412389363, 0.0453824653196, 0.0541722382861, 0.0638498928693, 0.0748241181818, 0.0870018575601, 0.100969984201, 0.116901498992, 0.135225153396, 0.155865741879, 0.179184255267, 0.205004290279, 0.233612351761, 0.265481420095, 0.307980863])
kf = np.array([188.0535454, 200.042702173, 212.031762644, 224.020823148, 236.009883356, 247.998943263, 259.988002683, 271.977061559, 283.96611989, 295.955177493, 307.944233912, 319.933288966, 331.922341978, 343.911392444, 355.900439242, 367.889482121, 379.878519645, 391.867551768, 403.856577188, 415.845593859, 427.834570885])
PHI_train = torch.FloatTensor(phi)
KF_train = torch.FloatTensor(kf)

def ludwig_func(x, a, b, c):
    return a+b*(x**c)

class Reg(torch.nn.Module):
    def __init__(self,n_feature,n_hidden,n_output):
        super(Reg, self).__init__()
        self.layer1 = torch.nn.Linear(1,16)
        self.layer2 = torch.nn.Linear(16,8)
        #self.layer4 = torch.nn.Linear(32,16)
        self.predict = torch.nn.Linear(8,3)

    def forward(self,x):
        x = F.relu(self.layer1(x))
        x = F.relu(self.layer2(x))
        x = self.predict(x)
        return x
reg = Reg(n_feature = 1, n_hidden = 16, n_output = 3)

optimizer = torch.optim.SGD(reg.parameters(),lr=0.01)
loss_func = torch.nn.MSELoss()

for epoch in range(200):
    PRM = reg(PHI_train) #Tensor
    prm = PRM.detach().numpy() #Array

    for i in range(len(ludwig_prm)): #length of prm = 21

        a = prm[i,0]
        b = prm[i,1]
        c = prm[i,2]
        #d = prm[i,3]
        preds = ludwig_func(phi,a,b,c)#Array
        np.set_printoptions(precision=10)
        preds = np.nan_to_num(preds)
        #print(preds)

        Preds = torch.FloatTensor(preds)#Tensor
        Preds[Preds == float('inf')] = -0.2e+10
        #print(Preds)

        loss = loss_func(Preds,KF_train) #Tensor
        loss[loss == float('inf')] = 0.1e+29
        loss = Variable(loss,requires_grad = True)

        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
    print('Epoch:',epoch+1,'MESloss:',loss.item())

我为此苦苦挣扎了很长时间，不知道为什么。在互联网上，我也找不到任何东西。太好了，有人可以帮忙。

Answer 1

您好，您的问题在于输入数组的定义，如果您尝试例如进行更改的话：

phi = Variable(torch.Tensor([[10.0], [9.0], [3.0], [2.0]]))
kf = Variable(torch.Tensor([[90.0], [80.0], [50.0], [30.0]]))

该程序将正常工作

我认为您必须重塑输入内容

Logistic回归，除以零的误差

1 个答案: