B站-《PyTorch深度学习实践》刘二大人 Otto Group Product Classification第9讲作业

import torch
import pandas as pd
import numpy as np
from torch.utils.data import Dataset
from torch.utils.data import DataLoader
import torch.nn.functional as F
from sklearn import preprocessing 


class OttoDataset(Dataset):
    def __init__(self,file_path):
        data = pd.read_csv(file_path,delimiter=',')
        le = preprocessing.LabelEncoder()
        data['target'] = le.fit_transform(data['target'])
        self.x_data = torch.tensor(data.iloc[:,1:-1].values,dtype=torch.float32)
        self.y_data = torch.tensor(data['target'].values,dtype=torch.long)
        self.len = data.shape[0]

    def __getitem__(self, index):
        return self.x_data[index], self.y_data[index]
    
    def __len__(self):
        return self.len
    
train_dataset = OttoDataset('otto_train.csv')
#print(train_dataset.y_data)
test_dataset = pd.read_csv('otto_test.csv')
#print(test_dataset)
test_x_data = torch.tensor(test_dataset.iloc[:,1:].values,dtype=torch.float32)
#print(test_x_data.shape)

train_data_loader = DataLoader(dataset=train_dataset,shuffle=True,batch_size=32,num_workers=2)
#print(train_dataset.y_data)

class Model(torch.nn.Module):
    def __init__(self):
        super().__init__()
        self.l1 = torch.nn.Linear(93,128)
        self.l2 = torch.nn.Linear(128,64)
        self.l3 = torch.nn.Linear(64,32)
        self.l4 = torch.nn.Linear(32,9)
        self.activate = torch.nn.ReLU()

    def forward(self,x):
        x = self.activate(self.l1(x))
        x = self.activate(self.l2(x))
        x = self.activate(self.l3(x))
        return self.l4(x)
    
model = Model()

criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(),lr=0.01,momentum=0.5)

def train(epoch):
    running_loss = 0
    for batch_idx, data in enumerate(train_data_loader,0):
        inputs, labels = data
        y_pred = model(inputs)
        loss = criterion(y_pred,labels)

        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
        running_loss += loss.item()

        if  batch_idx % 300 == 299:
            print('epoch= ',epoch,'bacth_idx= ',batch_idx,'loss= ',running_loss / 300)
            running_loss = 0

def test():
    with torch.no_grad():
        outputs = model(test_x_data)
        print(outputs)
        outputs_propability = F.softmax(outputs,dim=1).numpy()
        predictions_df = pd.DataFrame(outputs_propability, columns=['Class_1', 'Class_2', 'Class_3', 'Class_4', 'Class_5', 'Class_6', 'Class_7', 'Class_8', 'Class_9'])
        predictions_df.insert(0, 'id', test_dataset['id'])
        predictions_df.to_csv('predictions.csv', index=False)

        
    



if __name__ == '__main__':
    for epoch in range(100):
        train(epoch)
        test()

多分类任务中，使用Crossentropyloss的时候，训练数据中的标签需要为long整型数据，以支持类别索引。

最终分数：