Pytorch---使用Pytorch实现逻辑回归进行分类

一、代码中的数据集可以点击以下链接进行下载

百度网盘提取码：lala

二、代码运行环境

Pytorch-gpu==1.7.1
Python==3.7

三、数据集处理的代码如下所示

import pandas as pd
import torch


# 进行数据的读取 *** 作
def make_dataset():
    data = pd.read_csv(r'dataset/credit.csv', header=None)
    X = data.iloc[:, :-1]
    Y = data.iloc[:, -1].replace(-1, 0)
    X = torch.from_numpy(X.values).type(torch.float32)
    Y = torch.from_numpy(Y.values.reshape(-1, 1)).type(torch.float32)
    return X, Y


if __name__ == '__main__':
    x, y = make_dataset()
    print(x)
    print(y)

四、模型的构建代码如下所示

from torch import nn


# 进行模型的构建
def make_model():
    model = nn.Sequential(
        nn.Linear(in_features=15, out_features=1),
        nn.Sigmoid()
    )
    return model


if __name__ == '__main__':
    my_model = make_model()
    print(my_model)

五、模型的训练代码如下所示

from torch import nn
from data_loader import make_dataset
from model_loader import make_model
import torch
import tqdm

# 数据的读取
X, Y = make_dataset()

# 模型的加载
model = make_model()

# 定义训练时的配置
loss_fn = nn.BCELoss()
optimizer = torch.optim.Adam(model.parameters(), lr=0.0001)
batches = 16
num_of_batch = 653 // 16
epoches = 1000
train_epoches = tqdm.tqdm(range(epoches))

# 开始进行训练
for epoch in train_epoches:
    for i in range(num_of_batch):
        start = i * batches
        end = start + batches
        x = X[start:end]
        y = Y[start:end]
        y_pred = model(x)
        loss = loss_fn(y_pred, y)
        optimizer.zero_grad()  # 首先清空梯度
        loss.backward()  # 进行梯度计算
        optimizer.step()  # 根据梯度进行优化

# 模型的保存
torch.save(model.state_dict(), r'model_data\model.pth')

六、模型的预测代码如下所示

import torch
from data_loader import make_dataset
from model_loader import make_model

# 数据的读取
X, Y = make_dataset()

# 模型的加载
model = make_model()
model_state_dict = torch.load(r'model_data/model.pth')
model.load_state_dict(model_state_dict)

# 模型的预测
print('模型的参数情况如下：')
print(model.state_dict())  # 输出模型的参数
print('------------------------------------------------------------------------------')

# 进行模型的评估
print('模型的识别正确率为 {:14f}'.format(((model(X).data.numpy() > 0.5).astype('int') == Y.numpy()).mean()))

七、代码的运行结果如下所示