Kaggle的比赛题目home-credit-default-risk的代码

# import torch # import torch.nn as nn # import numpy as np # import pandas as pd # import torch.optim as optim # from torch.utils.data import DataLoader,Dataset # device = torch.device("cpu" if torch.cuda.is_available() else "cpu") # class mydata_train(Dataset): # def __init__(self,path): # data=np.array(pd.read_csv(path)) # self.len=data.shape[0] # self.X=torch.from_numpy(data[:,:-1]).float() # self.y=torch.from_numpy(data[:,-1]).float() # def __getitem__(self, index): # return self.X[index],self.y[index] # def __len__(self): # return self.len # class Model(nn.Module): # def __init__(self): # super(Model,self).__init__() # self.lr1=nn.Linear(20,64) # self.lr2=nn.Linear(64,10) # self.lr3=nn.Linear(10,2) # self.relu = nn.ReLU() # self.soft=nn.Softmax() # def forward(self,x): # x=self.relu(self.lr1(x.float())) # x=self.lr2(x) # x=self.soft(self.lr3(x)) # return x # def accuracy(predictions, labels): # pred = torch.max(predictions.data, 1)[1] # rights = pred.eq(labels.data.view_as(pred)).sum() # return rights, len(labels) # train_dataset=mydata_train("torch_train.csv") # test_dataset=mydata_train("torch_test.csv") # train_loader=DataLoader(dataset=train_dataset,batch_size=128,shuffle=True) # test_loader=DataLoader(test_dataset,batch_size=64,shuffle=False) # model=Model() # criterion=nn.MSELoss() # optimizer=optim.Adam(model.parameters(),lr=0.1) # epochs=100 # for epoch in range(epochs): # model.train() # train_rights = [] # for id,(data,target) in enumerate(train_loader): # data=data.to(device) # target=target.to(device) # target = torch.unsqueeze(target, dim=1) # optimizer.zero_grad() # output=model(data) # loss=criterion(output,target) # loss.backward() # optimizer.step() # right = accuracy(output, target) # train_rights.append(right) # model.eval() # test_rights=[] # for id,(data,target) in enumerate(test_loader): # data = data.to(device) # target = target.to(device) # output = model(data) # right = accuracy(output, target) # test_rights.append(right) # train_r = (sum([tup[0] for tup in train_rights]), sum([tup[1] for tup in train_rights])) # test_r = (sum([tup[0] for tup in test_rights]), sum([tup[1] for tup in test_rights])) # print(f"train_acc:{(100.*train_r[0]/train_r[1]).item()},test_acc:{(100.*test_r[0]/test_r[1]).item()}") import torch import torch.nn as nn import numpy as np import pandas as pd import torch.optim as optim from torch.utils.data import DataLoader, Dataset device = torch.device("cpu" if torch.cuda.is_available() else "cpu") class mydata_train(Dataset): def __init__(self, path): data = np.array(pd.read_csv(path)) self.len = data.shape[0] self.X = torch.from_numpy(data[:, :-1]).float() self.y = torch.from_numpy(data[:, -1]).long() def __getitem__(self, index): return self.X[index], self.y[index] def __len__(self): return self.len class Model(nn.Module): def __init__(self): super(Model, self).__init__() self.lr1 = nn.Linear(20, 64) self.lr2 = nn.Linear(64, 10) self.lr3 = nn.Linear(10, 2) self.relu = nn.ReLU() self.soft = nn.LogSoftmax(dim=1) def forward(self, x): x = self.relu(self.lr1(x.float())) x = self.lr2(x) x = self.soft(self.lr3(x)) return x def accuracy(predictions, labels): pred = torch.max(predictions, 1)[1] rights = pred.eq(labels.data.view_as(pred)).sum() return rights, len(labels) train_dataset = mydata_train("torch_train.csv") test_dataset = mydata_train("torch_test.csv") train_loader = DataLoader(dataset=train_dataset, batch_size=128, shuffle=True) test_loader = DataLoader(test_dataset, batch_size=64, shuffle=False) model = Model() criterion = nn.NLLLoss() optimizer = optim.Adam(model.parameters(), lr=0.1) epochs = 100 for epoch in range(epochs): model.train() train_rights = [] for id, (data, target) in enumerate(train_loader): data = data.to(device) target = target.to(device) optimizer.zero_grad() output = model(data) loss = criterion(output, target) loss.backward() optimizer.step() right = accuracy(output, target) train_rights.append(right) model.eval() test_rights = [] for id, (data, target) in enumerate(test_loader): data = data.to(device) target = target.to(device) output = model(data) right = accuracy(output, target) test_rights.append(right) train_r = (sum([tup[0] for tup in train_rights]), sum([tup[1] for tup in train_rights])) test_r = (sum([tup[0] for tup in test_rights]), sum([tup[1] for tup in test_rights])) print(f"第{epoch}次：train_acc: {(100. * train_r[0] / train_r[1]).item()}, test_acc: {(100. * test_r[0] / test_r[1]).item()}")