代碼主要核心思想來自：https://www.cnblogs.com/JadenFK3326/p/12164519.html

K折交叉交叉驗證的過程如下：

以200條數(shù)據(jù)，十折交叉驗證為例子，十折也就是將數(shù)據(jù)分成10組，進行10組訓練，每組用于測試的數(shù)據(jù)為：數(shù)據(jù)總條數(shù)/組數(shù)，即每組20條用于valid，180條用于train，每次valid的都是不同的。

（1）將200條數(shù)據(jù)，分成按照 數(shù)據(jù)總條數(shù)/組數(shù)（折數(shù)），進行切分。然后取出第i份作為第i次的valid，剩下的作為train

（2）將每組中的train數(shù)據(jù)利用DataLoader和Dataset，進行封裝。

（3）將train數(shù)據(jù)用于訓練，epoch可以自己定義，然后利用valid做驗證。得到一次的train_loss和?valid_loss。

（4）重復（2）（3）步驟，得到最終的 averge_train_loss和averge_valid_loss

上述過程如下圖所示：

上述的代碼如下：

import torch import torch.nn as nn from torch.utils.data import DataLoader,Dataset import torch.nn.functional as F from torch.autograd import Variable#####構造的訓練集#### x = torch.rand(100,28,28) y = torch.randn(100,28,28) x = torch.cat((x,y),dim=0) label =[1] *100 + [0]*100 label = torch.tensor(label,dtype=torch.long)######網(wǎng)絡結構########## class Net(nn.Module):#定義Netdef __init__(self):super(Net, self).__init__() self.fc1 = nn.Linear(28*28, 120) self.fc2 = nn.Linear(120, 84)self.fc3 = nn.Linear(84, 2)def forward(self, x):x = x.view(-1, self.num_flat_features(x)) x = F.relu(self.fc1(x)) x = F.relu(self.fc2(x)) x = self.fc3(x) return xdef num_flat_features(self, x):size = x.size()[1:] num_features = 1for s in size:num_features *= sreturn num_features##########定義dataset########## class TraindataSet(Dataset):def __init__(self,train_features,train_labels):self.x_data = train_featuresself.y_data = train_labelsself.len = len(train_labels)def __getitem__(self,index):return self.x_data[index],self.y_data[index]def __len__(self):return self.len########k折劃分############ def get_k_fold_data(k, i, X, y): ###此過程主要是步驟（1）# 返回第i折交叉驗證時所需要的訓練和驗證數(shù)據(jù)，分開放，X_train為訓練數(shù)據(jù)，X_valid為驗證數(shù)據(jù)assert k > 1fold_size = X.shape[0] // k # 每份的個數(shù):數(shù)據(jù)總條數(shù)/折數(shù)（組數(shù)）X_train, y_train = None, Nonefor j in range(k):idx = slice(j * fold_size, (j + 1) * fold_size) #slice(start,end,step)切片函數(shù)##idx 為每組 validX_part, y_part = X[idx, :], y[idx]if j == i: ###第i折作validX_valid, y_valid = X_part, y_partelif X_train is None:X_train, y_train = X_part, y_partelse:X_train = torch.cat((X_train, X_part), dim=0) #dim=0增加行數(shù)，豎著連接y_train = torch.cat((y_train, y_part), dim=0)#print(X_train.size(),X_valid.size())return X_train, y_train, X_valid,y_validdef k_fold(k, X_train, y_train, num_epochs=3,learning_rate=0.001, weight_decay=0.1, batch_size=5):train_loss_sum, valid_loss_sum = 0, 0train_acc_sum ,valid_acc_sum = 0,0for i in range(k):data = get_k_fold_data(k, i, X_train, y_train) # 獲取k折交叉驗證的訓練和驗證數(shù)據(jù)net = Net() ### 實例化模型### 每份數(shù)據(jù)進行訓練,體現(xiàn)步驟三####train_ls, valid_ls = train(net, *data, num_epochs, learning_rate,\weight_decay, batch_size) print('*'*25,'第',i+1,'折','*'*25)print('train_loss:%.6f'%train_ls[-1][0],'train_acc:%.4f\n'%valid_ls[-1][1],\'valid loss:%.6f'%valid_ls[-1][0],'valid_acc:%.4f'%valid_ls[-1][1])train_loss_sum += train_ls[-1][0]valid_loss_sum += valid_ls[-1][0]train_acc_sum += train_ls[-1][1]valid_acc_sum += valid_ls[-1][1]print('#'*10,'最終k折交叉驗證結果','#'*10) ####體現(xiàn)步驟四#####print('train_loss_sum:%.4f'%(train_loss_sum/k),'train_acc_sum:%.4f\n'%(train_acc_sum/k),\'valid_loss_sum:%.4f'%(valid_loss_sum/k),'valid_acc_sum:%.4f'%(valid_acc_sum/k))#########訓練函數(shù)########## def train(net, train_features, train_labels, test_features, test_labels, num_epochs, learning_rate,weight_decay, batch_size):train_ls, test_ls = [], [] ##存儲train_loss,test_lossdataset = TraindataSet(train_features, train_labels) train_iter = DataLoader(dataset, batch_size, shuffle=True) ### 將數(shù)據(jù)封裝成 Dataloder 對應步驟（2）#這里使用了Adam優(yōu)化算法optimizer = torch.optim.Adam(params=net.parameters(), lr= learning_rate, weight_decay=weight_decay)for epoch in range(num_epochs):for X, y in train_iter: ###分批訓練 output = net(X)loss = loss_func(output,y)optimizer.zero_grad()loss.backward()optimizer.step()### 得到每個epoch的 loss 和 accuracy train_ls.append(log_rmse(0,net, train_features, train_labels)) if test_labels is not None:test_ls.append(log_rmse(1,net, test_features, test_labels))#print(train_ls,test_ls)return train_ls, test_lsdef log_rmse(flag,net,x,y):if flag == 1: ### valid 數(shù)據(jù)集net.eval()output = net(x)result = torch.max(output,1)[1].view(y.size())corrects = (result.data == y.data).sum().item()accuracy = corrects*100.0/len(y) #### 5 是 batch_sizeloss = loss_func(output,y)net.train()return (loss.data.item(),accuracy)loss_func = nn.CrossEntropyLoss() ###申明loss函 k_fold(10,x,label) ### k=10,十折交叉驗證

上述代碼中，直接按照順序從x中每次截取20條作為valid，也可以先打亂然后在截取，這樣效果應該會更好。如下所示：

import random import torchx = torch.rand(100,28,28) y = torch.randn(100,28,28) x = torch.cat((x,y),dim=0) label =[1] *100 + [0]*100 label = torch.tensor(label,dtype=torch.long)index = [i for i in range(len(x))] random.shuffle(index) x = x[index] label = label[index]

?

總結

以上是生活随笔為你收集整理的pytorch - K折交叉验证过程说明及实现的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯，歡迎將生活随笔推薦給好友。