基于RNN模型的心脏病预测，提供tensorflow和pytorch实现

• 🍨 本文为🔗365天深度学习训练营 中的学习记录博客
• 🍖 原作者：K同学啊

前言

• RNN是很经典的模型，原理参考：深度学习基础--一文搞懂RNN

• 这个案例是一个基础案例，用RNN模型去做一个二分类问题，心脏病预测，数据集在kaggle上可以找到；

• 本篇为pytorch实现，TensorFlow实现为：基于RNN模型的心脏病预测(tensorflow实现)

• 欢迎收藏加关注，本人将会持续更新。

  文章目录

  • • 1、数据处理
    • • 1、导入库
      • 2、导入数据
      • 3、数据分析
      • • 数据初步分析
        • 缺失值
        • 相关性分析
      • 4、数据划分
      • 5、数据标准化
      • 6、转化为张量数据
    • 2、创建模型
    • 3、模型训练
    • • 1、设置超参数
      • 2、设置训练函数
      • 3、设置测试函数
    • 4、模型训练
    • 5、结果展示
    • 6、模型评估

1、数据处理

1、导入库

import pandas as pd 
import numpy as np 
import matplotlib.pyplot as plt 
from torch.utils.data import DataLoader, TensorDataset
import torch 


device = 'cuda' if torch.cuda.is_available() else 'cpu'
device

'cuda'

2、导入数据

data = pd.read_csv('./heart.csv')

data.head()

| | age | sex | cp | trestbps | chol | fbs | restecg | thalach | exang | oldpeak | slope | ca | thal | target |
| 0 | 63 | 1 | 3 | 145 | 233 | 1 | 0 | 150 | 0 | 2.3 | 0 | 0 | 1 | 1 |
| 1 | 37 | 1 | 2 | 130 | 250 | 0 | 1 | 187 | 0 | 3.5 | 0 | 0 | 2 | 1 |
| 2 | 41 | 0 | 1 | 130 | 204 | 0 | 0 | 172 | 0 | 1.4 | 2 | 0 | 2 | 1 |
| 3 | 56 | 1 | 1 | 120 | 236 | 0 | 1 | 178 | 0 | 0.8 | 2 | 0 | 2 | 1 |

4	57	0	0	120	354	0	1	163	1	0.6	2	0	2	1

• age - 年龄
• sex - (1 = male(男性); 0 = (女性))
• cp - chest pain type(胸部疼痛类型)（1：典型的心绞痛-typical，2：非典型心绞痛-atypical，3：没有心绞痛-non-anginal，4：无症状-asymptomatic）
• trestbps - 静息血压 (in mm Hg on admission to the hospital)
• chol - 胆固醇 in mg/dl
• fbs - (空腹血糖 > 120 mg/dl) (1 = true; 0 = false)
• restecg - 静息心电图测量（0：普通，1：ST-T波异常，2：可能左心室肥大）
• thalach - 最高心跳率
• exang - 运动诱发心绞痛 (1 = yes; 0 = no)
• oldpeak - 运动相对于休息引起的ST抑制
• slope - 运动ST段的峰值斜率（1：上坡-upsloping，2：平的-flat，3：下坡-downsloping）
• ca - 主要血管数目(0-4)
• thal - 一种叫做地中海贫血的血液疾病（3 = normal; 6 = 固定的缺陷-fixed defect; 7 = 可逆的缺陷-reversable defect）
• target - 是否患病 (1=yes, 0=no)

3、数据分析

数据初步分析

data.info()   # 数据类型分析

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 303 entries, 0 to 302
Data columns (total 14 columns):
 #   Column    Non-Null Count  Dtype  
---  ------    --------------  -----  
 0   age       303 non-null    int64  
 1   sex       303 non-null    int64  
 2   cp        303 non-null    int64  
 3   trestbps  303 non-null    int64  
 4   chol      303 non-null    int64  
 5   fbs       303 non-null    int64  
 6   restecg   303 non-null    int64  
 7   thalach   303 non-null    int64  
 8   exang     303 non-null    int64  
 9   oldpeak   303 non-null    float64
 10  slope     303 non-null    int64  
 11  ca        303 non-null    int64  
 12  thal      303 non-null    int64  
 13  target    303 non-null    int64  
dtypes: float64(1), int64(13)
memory usage: 33.3 KB

其中分类变量为：sex、cp、fbs、restecg、exang、slope、ca、thal、target

数值型变量：age、trestbps、chol、thalach、oldpeak

data.describe()  # 描述性

| | age | sex | cp | trestbps | chol | fbs | restecg | thalach | exang | oldpeak | slope | ca | thal | target |
| count | 303.000000 | 303.000000 | 303.000000 | 303.000000 | 303.000000 | 303.000000 | 303.000000 | 303.000000 | 303.000000 | 303.000000 | 303.000000 | 303.000000 | 303.000000 | 303.000000 |
| mean | 54.366337 | 0.683168 | 0.966997 | 131.623762 | 246.264026 | 0.148515 | 0.528053 | 149.646865 | 0.326733 | 1.039604 | 1.399340 | 0.729373 | 2.313531 | 0.544554 |
| std | 9.082101 | 0.466011 | 1.032052 | 17.538143 | 51.830751 | 0.356198 | 0.525860 | 22.905161 | 0.469794 | 1.161075 | 0.616226 | 1.022606 | 0.612277 | 0.498835 |
| min | 29.000000 | 0.000000 | 0.000000 | 94.000000 | 126.000000 | 0.000000 | 0.000000 | 71.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 |
| 25% | 47.500000 | 0.000000 | 0.000000 | 120.000000 | 211.000000 | 0.000000 | 0.000000 | 133.500000 | 0.000000 | 0.000000 | 1.000000 | 0.000000 | 2.000000 | 0.000000 |
| 50% | 55.000000 | 1.000000 | 1.000000 | 130.000000 | 240.000000 | 0.000000 | 1.000000 | 153.000000 | 0.000000 | 0.800000 | 1.000000 | 0.000000 | 2.000000 | 1.000000 |
| 75% | 61.000000 | 1.000000 | 2.000000 | 140.000000 | 274.500000 | 0.000000 | 1.000000 | 166.000000 | 1.000000 | 1.600000 | 2.000000 | 1.000000 | 3.000000 | 1.000000 |

max	77.000000	1.000000	3.000000	200.000000	564.000000	1.000000	2.000000	202.000000	1.000000	6.200000	2.000000	4.000000	3.000000	1.000000

• 年纪：均值54，中位数55，标准差9，说明主要是老年人，偏大
• 静息血压：均值131.62， 成年人一般：正常血压：收缩压 < 120 mmHg，偏大
• 胆固醇：均值246.26，理想水平：小于 200 mg/dL，偏大
• 最高心率：均值149.64，一般静息状态下通常是 60 到 100 次每分钟，偏大

最大值和最小值都可能发生，无异常值

缺失值

data.isnull().sum()

age         0
sex         0
cp          0
trestbps    0
chol        0
fbs         0
restecg     0
thalach     0
exang       0
oldpeak     0
slope       0
ca          0
thal        0
target      0
dtype: int64

相关性分析

import seaborn as sns

plt.figure(figsize=(20, 15))

sns.heatmap(data.corr(), annot=True, cmap='Greens')

plt.show()

​

​

相关系数的等级划分

• 非常弱的相关性：
  • 0.00 至 0.19 或 -0.00 至 -0.19
  • 解释：几乎不存在线性关系。
• 弱相关性：
  • 0.20 至 0.39 或 -0.20 至 -0.39
  • 解释：存在一定的线性关系，但较弱。
• 中等相关性：
  • 0.40 至 0.59 或 -0.40 至 -0.59
  • 解释：有明显的线性关系，但不是特别强。
• 强相关性：
  • 0.60 至 0.79 或 -0.60 至 -0.79
  • 解释：两个变量之间有较强的线性关系。
• 非常强的相关性：
  • 0.80 至 1.00 或 -0.80 至 -1.00
  • 解释：几乎完全线性相关，表明两个变量的变化高度一致。

---

target与chol、没有什么相关性，fbs是分类变量，chol胆固醇是数值型变量，但是从实际角度，这些都有影响，故不剔除特征

4、数据划分

这里先划分为：训练集：测试集 = 9：1

from sklearn.model_selection import train_test_split

X = data.iloc[:, :-1]
y = data.iloc[:, -1]

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.1, random_state=42)

5、数据标准化

from sklearn.preprocessing import StandardScaler

scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

# 深度学习、用rnn模型，数据需要3通道，在图片中表示RGB，这里表示1
X_train = X_train.reshape(X_train.shape[0], X_train.shape[1], 1)
X_test = X_test.reshape(X_test.shape[0], X_test.shape[1], 1)

6、转化为张量数据

# 假设  y_train, y_test 是 pandas Series 或 DataFrame
# 首先将它们转换为 NumPy 数组
y_train = y_train.values.astype(np.float32)
y_test = y_test.values.astype(np.float32)

batch_size = 32

# unsqueeze  (N,) 转换为 (N, 1)
train_dataset = TensorDataset(torch.tensor(X_train, dtype=torch.float32).to(device), torch.tensor(y_train, dtype=torch.float32).unsqueeze(1).to(device)) 
test_dataset = TensorDataset(torch.tensor(X_test, dtype=torch.float32).to(device), torch.tensor(y_test, dtype=torch.float32).unsqueeze(1).to(device))

train_dl = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)
test_dl = DataLoader(test_dataset, batch_size=batch_size, shuffle=True)

2、创建模型

• 定义一个RNN层
  rnn = nn.RNN(input_size=10, hidden_size=20, num_layers=2, nonlinearity='tanh',
  bias=True, batch_first=False, dropout=0, bidirectional=False)
• input_size: 输入的特征维度
• hidden_size: 隐藏层的特征维度
• num_layers: RNN 层的数量
• nonlinearity: 非线性激活函数 ('tanh' 或 'relu')
• bias: 如果为 False，则内部不含偏置项，默认为 True
• batch_first: 如果为 True，则输入和输出张量提供为 (batch, seq, feature)，默认为 False (seq, batch, feature)
• dropout: 如果非零，则除了最后一层，在每层的输出中引入一个 Dropout 层，默认为 0
• bidirectional: 如果为 True，则将成为双向 RNN，默认为 False

import torch  
import torch.nn as nn 

# 创建模型
'''
该问题本质是二分类问题，故最后一层全连接层用激活函数为：sigmoid
模型结构：
    RNN：隐藏层200，激活函数：relu
    Linear：--> 100(relu) -> 1(sigmoid)
'''
# 创建模型
class Model(nn.Module):
    def __init__(self):
        super().__init__()
        
        # 在 Keras 中 input_shape=(13, 1) 表示的是, 每个样本有 13 个时间步（seq_length=13），每个时间步有一个特征（input_size=1）, 换句话就是一行
        self.rnn = nn.RNN(input_size=1, hidden_size=200, num_layers=1, nonlinearity='relu', batch_first=True)
        
        self.fc1 = nn.Linear(200, 100)
        self.fc2 = nn.Linear(100, 1)
        
    def forward(self, x):
        # 初始化隐藏层状态
        h0 = torch.zeros(1, x.size(0), 200).to(device)  # (num_layers, batch_size, hidden_size)
        # 构建神经网络
        x, _ = self.rnn(x, h0)  # x: (batch_size, seq_length, hidden_size)
        x = x[:, -1, :] # 最后一个时间步作为全连接层的输入, 形状变为：(batch_size, input_size)
        x = torch.relu(self.fc1(x))
        x = torch.sigmoid(self.fc2(x))
        return x
    

model = Model().to(device)
        

3、模型训练

1、设置超参数

loss_fn = nn.BCELoss()
lr = 1e-4
optimizer = torch.optim.Adam(model.parameters(), lr=lr)

2、设置训练函数

def train(dataloader, model, loss_fn, optimizer):
    # 总大小
    size = len(dataloader.dataset)
    # 批次数量
    num_batches = len(dataloader)

    # 准确率和损失初始化
    correct = 0
    running_loss = 0.0

    for X, y in dataloader:
        X, y = X.to(device), y.to(device)

        # 模型预测与误差评分
        pred = model(X).squeeze()  # 去除多余的维度以匹配目标形状
        if y.dim() == 2:  # 如果目标形状是 [batch_size, 1]
            y = y.squeeze()  # 将其转换为 [batch_size]
        loss = loss_fn(pred, y)  # 确保目标形状匹配
        
        # 梯度清零
        optimizer.zero_grad()
        
        # 反向传播与梯度更新
        loss.backward()
        optimizer.step()

        # 记录损失
        running_loss += loss.item()

        # 计算准确率, 二分类和多分类不同
        predicted_labels = (pred > 0.5).float()  # 使用 0.5 作为阈值
        correct += (predicted_labels == y).type(torch.float64).sum().item()

    # 计算平均损失和准确率
    train_loss = running_loss / num_batches
    train_acc = correct / size  

    return train_acc, train_loss

3、设置测试函数

def test(dataloader, model, loss_fn):
    # 总大小
    size = len(dataloader.dataset)
    # 批次数量
    num_batches = len(dataloader)

    # 准确率和损失初始化
    correct = 0
    running_loss = 0.0

    # 将模型设置为评估模式
    model.eval()

    with torch.no_grad():
        for X, y in dataloader:
            X, y = X.to(device), y.to(device)

            # 模型预测与误差评分
            pred = model(X).squeeze()  # 去除多余的维度以匹配目标形状
            if y.dim() == 2:  # 如果目标形状是 [batch_size, 1]
                y = y.squeeze()  # 将其转换为 [batch_size]
            loss = loss_fn(pred, y)  # 确保目标形状匹配

            # 记录损失
            running_loss += loss.item()

            # 计算准确率
            predicted_labels = (pred > 0.5).float()  # 使用 0.5 作为阈值
            correct += (predicted_labels == y).type(torch.float64).sum().item()

    # 计算平均损失和准确率
    test_loss = running_loss / num_batches
    test_acc = correct / size  # 转换为百分比

    return test_acc, test_loss

4、模型训练

train_acc = []
train_loss = []
test_acc = []
test_loss = []

# 定义训练次数
epoches = 100

for epoch in range(epoches):
    # 训练
    model.train()
    epoch_trian_acc, epoch_train_loss = train(train_dl, model, loss_fn, optimizer)
    
    # 测试
    model.eval()
    epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)
    
    # 记录
    train_acc.append(epoch_trian_acc)
    train_loss.append(epoch_train_loss)
    test_acc.append(epoch_test_acc)
    test_loss.append(epoch_test_loss)
    
    template = ('Epoch:{:2d}, Train_acc:{:.1f}%, Train_loss:{:.3f}, Test_acc:{:.1f}%, Test_loss:{:.3f}')
    print(template.format(epoch+1, epoch_trian_acc*100, epoch_train_loss, epoch_test_acc*100, epoch_test_loss))

Epoch: 1, Train_acc:71.0%, Train_loss:0.689, Test_acc:71.0%, Test_loss:0.688
Epoch: 2, Train_acc:75.0%, Train_loss:0.686, Test_acc:71.0%, Test_loss:0.685
Epoch: 3, Train_acc:75.0%, Train_loss:0.682, Test_acc:71.0%, Test_loss:0.682
Epoch: 4, Train_acc:75.0%, Train_loss:0.678, Test_acc:71.0%, Test_loss:0.678
Epoch: 5, Train_acc:75.0%, Train_loss:0.673, Test_acc:71.0%, Test_loss:0.674
Epoch: 6, Train_acc:75.0%, Train_loss:0.669, Test_acc:71.0%, Test_loss:0.670
Epoch: 7, Train_acc:75.0%, Train_loss:0.661, Test_acc:71.0%, Test_loss:0.664
Epoch: 8, Train_acc:75.4%, Train_loss:0.657, Test_acc:67.7%, Test_loss:0.656
Epoch: 9, Train_acc:77.2%, Train_loss:0.644, Test_acc:67.7%, Test_loss:0.647
Epoch:10, Train_acc:77.6%, Train_loss:0.635, Test_acc:71.0%, Test_loss:0.632
Epoch:11, Train_acc:79.0%, Train_loss:0.615, Test_acc:74.2%, Test_loss:0.613
Epoch:12, Train_acc:79.0%, Train_loss:0.592, Test_acc:77.4%, Test_loss:0.585
Epoch:13, Train_acc:80.5%, Train_loss:0.559, Test_acc:77.4%, Test_loss:0.557
Epoch:14, Train_acc:77.9%, Train_loss:0.536, Test_acc:77.4%, Test_loss:0.532
Epoch:15, Train_acc:78.7%, Train_loss:0.508, Test_acc:74.2%, Test_loss:0.520
Epoch:16, Train_acc:77.9%, Train_loss:0.490, Test_acc:77.4%, Test_loss:0.510
Epoch:17, Train_acc:79.4%, Train_loss:0.482, Test_acc:74.2%, Test_loss:0.510
Epoch:18, Train_acc:79.0%, Train_loss:0.459, Test_acc:74.2%, Test_loss:0.505
Epoch:19, Train_acc:80.9%, Train_loss:0.440, Test_acc:74.2%, Test_loss:0.513
Epoch:20, Train_acc:79.8%, Train_loss:0.426, Test_acc:74.2%, Test_loss:0.522
Epoch:21, Train_acc:78.7%, Train_loss:0.424, Test_acc:74.2%, Test_loss:0.529
Epoch:22, Train_acc:77.6%, Train_loss:0.447, Test_acc:71.0%, Test_loss:0.538
Epoch:23, Train_acc:79.0%, Train_loss:0.441, Test_acc:74.2%, Test_loss:0.553
Epoch:24, Train_acc:80.5%, Train_loss:0.400, Test_acc:74.2%, Test_loss:0.517
Epoch:25, Train_acc:80.9%, Train_loss:0.421, Test_acc:74.2%, Test_loss:0.522
Epoch:26, Train_acc:80.1%, Train_loss:0.396, Test_acc:77.4%, Test_loss:0.539
Epoch:27, Train_acc:79.8%, Train_loss:0.393, Test_acc:77.4%, Test_loss:0.525
Epoch:28, Train_acc:81.2%, Train_loss:0.390, Test_acc:77.4%, Test_loss:0.524
Epoch:29, Train_acc:80.1%, Train_loss:0.378, Test_acc:77.4%, Test_loss:0.543
Epoch:30, Train_acc:80.1%, Train_loss:0.384, Test_acc:80.6%, Test_loss:0.521
Epoch:31, Train_acc:82.0%, Train_loss:0.392, Test_acc:77.4%, Test_loss:0.534
Epoch:32, Train_acc:81.6%, Train_loss:0.371, Test_acc:77.4%, Test_loss:0.513
Epoch:33, Train_acc:83.5%, Train_loss:0.376, Test_acc:77.4%, Test_loss:0.526
Epoch:34, Train_acc:81.6%, Train_loss:0.365, Test_acc:80.6%, Test_loss:0.511
Epoch:35, Train_acc:82.0%, Train_loss:0.383, Test_acc:77.4%, Test_loss:0.521
Epoch:36, Train_acc:83.8%, Train_loss:0.362, Test_acc:80.6%, Test_loss:0.513
Epoch:37, Train_acc:83.8%, Train_loss:0.357, Test_acc:80.6%, Test_loss:0.511
Epoch:38, Train_acc:84.2%, Train_loss:0.360, Test_acc:80.6%, Test_loss:0.511
Epoch:39, Train_acc:84.2%, Train_loss:0.354, Test_acc:80.6%, Test_loss:0.503
Epoch:40, Train_acc:84.9%, Train_loss:0.349, Test_acc:80.6%, Test_loss:0.512
Epoch:41, Train_acc:84.6%, Train_loss:0.371, Test_acc:80.6%, Test_loss:0.503
Epoch:42, Train_acc:84.6%, Train_loss:0.338, Test_acc:80.6%, Test_loss:0.510
Epoch:43, Train_acc:83.5%, Train_loss:0.353, Test_acc:80.6%, Test_loss:0.503
Epoch:44, Train_acc:83.8%, Train_loss:0.351, Test_acc:80.6%, Test_loss:0.500
Epoch:45, Train_acc:84.6%, Train_loss:0.339, Test_acc:80.6%, Test_loss:0.505
Epoch:46, Train_acc:85.7%, Train_loss:0.336, Test_acc:80.6%, Test_loss:0.500
Epoch:47, Train_acc:84.6%, Train_loss:0.358, Test_acc:80.6%, Test_loss:0.503
Epoch:48, Train_acc:84.9%, Train_loss:0.337, Test_acc:80.6%, Test_loss:0.513
Epoch:49, Train_acc:86.0%, Train_loss:0.334, Test_acc:80.6%, Test_loss:0.497
Epoch:50, Train_acc:85.3%, Train_loss:0.341, Test_acc:77.4%, Test_loss:0.513
Epoch:51, Train_acc:84.9%, Train_loss:0.337, Test_acc:80.6%, Test_loss:0.498
Epoch:52, Train_acc:84.9%, Train_loss:0.340, Test_acc:80.6%, Test_loss:0.499
Epoch:53, Train_acc:86.4%, Train_loss:0.328, Test_acc:80.6%, Test_loss:0.497
Epoch:54, Train_acc:84.9%, Train_loss:0.331, Test_acc:80.6%, Test_loss:0.502
Epoch:55, Train_acc:84.2%, Train_loss:0.343, Test_acc:77.4%, Test_loss:0.521
Epoch:56, Train_acc:84.6%, Train_loss:0.346, Test_acc:80.6%, Test_loss:0.486
Epoch:57, Train_acc:85.3%, Train_loss:0.351, Test_acc:77.4%, Test_loss:0.506
Epoch:58, Train_acc:85.7%, Train_loss:0.317, Test_acc:80.6%, Test_loss:0.491
Epoch:59, Train_acc:84.9%, Train_loss:0.327, Test_acc:77.4%, Test_loss:0.502
Epoch:60, Train_acc:86.0%, Train_loss:0.321, Test_acc:80.6%, Test_loss:0.503
Epoch:61, Train_acc:87.1%, Train_loss:0.340, Test_acc:80.6%, Test_loss:0.498
Epoch:62, Train_acc:85.3%, Train_loss:0.319, Test_acc:77.4%, Test_loss:0.501
Epoch:63, Train_acc:86.0%, Train_loss:0.317, Test_acc:77.4%, Test_loss:0.503
Epoch:64, Train_acc:86.4%, Train_loss:0.315, Test_acc:80.6%, Test_loss:0.493
Epoch:65, Train_acc:86.0%, Train_loss:0.323, Test_acc:80.6%, Test_loss:0.499
Epoch:66, Train_acc:86.8%, Train_loss:0.322, Test_acc:77.4%, Test_loss:0.518
Epoch:67, Train_acc:87.1%, Train_loss:0.308, Test_acc:80.6%, Test_loss:0.494
Epoch:68, Train_acc:86.8%, Train_loss:0.335, Test_acc:80.6%, Test_loss:0.507
Epoch:69, Train_acc:86.4%, Train_loss:0.307, Test_acc:80.6%, Test_loss:0.499
Epoch:70, Train_acc:86.4%, Train_loss:0.306, Test_acc:80.6%, Test_loss:0.505
Epoch:71, Train_acc:86.0%, Train_loss:0.314, Test_acc:77.4%, Test_loss:0.510
Epoch:72, Train_acc:86.8%, Train_loss:0.315, Test_acc:80.6%, Test_loss:0.495
Epoch:73, Train_acc:86.0%, Train_loss:0.311, Test_acc:77.4%, Test_loss:0.507
Epoch:74, Train_acc:86.8%, Train_loss:0.308, Test_acc:77.4%, Test_loss:0.512
Epoch:75, Train_acc:86.0%, Train_loss:0.316, Test_acc:80.6%, Test_loss:0.497
Epoch:76, Train_acc:85.7%, Train_loss:0.311, Test_acc:80.6%, Test_loss:0.504
Epoch:77, Train_acc:86.8%, Train_loss:0.307, Test_acc:77.4%, Test_loss:0.505
Epoch:78, Train_acc:86.4%, Train_loss:0.303, Test_acc:77.4%, Test_loss:0.508
Epoch:79, Train_acc:87.5%, Train_loss:0.296, Test_acc:80.6%, Test_loss:0.507
Epoch:80, Train_acc:87.1%, Train_loss:0.310, Test_acc:80.6%, Test_loss:0.508
Epoch:81, Train_acc:87.5%, Train_loss:0.297, Test_acc:77.4%, Test_loss:0.503
Epoch:82, Train_acc:87.5%, Train_loss:0.288, Test_acc:77.4%, Test_loss:0.527
Epoch:83, Train_acc:87.1%, Train_loss:0.293, Test_acc:80.6%, Test_loss:0.502
Epoch:84, Train_acc:87.1%, Train_loss:0.295, Test_acc:80.6%, Test_loss:0.508
Epoch:85, Train_acc:87.1%, Train_loss:0.283, Test_acc:80.6%, Test_loss:0.509
Epoch:86, Train_acc:87.1%, Train_loss:0.282, Test_acc:77.4%, Test_loss:0.514
Epoch:87, Train_acc:87.5%, Train_loss:0.278, Test_acc:80.6%, Test_loss:0.511
Epoch:88, Train_acc:87.5%, Train_loss:0.287, Test_acc:80.6%, Test_loss:0.513
Epoch:89, Train_acc:88.6%, Train_loss:0.308, Test_acc:77.4%, Test_loss:0.521
Epoch:90, Train_acc:87.9%, Train_loss:0.296, Test_acc:80.6%, Test_loss:0.512
Epoch:91, Train_acc:87.5%, Train_loss:0.287, Test_acc:77.4%, Test_loss:0.522
Epoch:92, Train_acc:87.5%, Train_loss:0.285, Test_acc:80.6%, Test_loss:0.512
Epoch:93, Train_acc:87.9%, Train_loss:0.287, Test_acc:80.6%, Test_loss:0.512
Epoch:94, Train_acc:88.2%, Train_loss:0.280, Test_acc:77.4%, Test_loss:0.530
Epoch:95, Train_acc:88.6%, Train_loss:0.283, Test_acc:80.6%, Test_loss:0.512
Epoch:96, Train_acc:89.3%, Train_loss:0.280, Test_acc:80.6%, Test_loss:0.516
Epoch:97, Train_acc:87.9%, Train_loss:0.276, Test_acc:77.4%, Test_loss:0.514
Epoch:98, Train_acc:88.6%, Train_loss:0.270, Test_acc:77.4%, Test_loss:0.526
Epoch:99, Train_acc:89.0%, Train_loss:0.269, Test_acc:80.6%, Test_loss:0.517
Epoch:100, Train_acc:88.6%, Train_loss:0.266, Test_acc:80.6%, Test_loss:0.521

5、结果展示

import matplotlib.pyplot as plt
#隐藏警告
import warnings
warnings.filterwarnings("ignore")               #忽略警告信息
plt.rcParams['font.sans-serif']    = ['SimHei'] # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False      # 用来正常显示负号
plt.rcParams['figure.dpi']         = 100        #分辨率

epoch_length = range(epoches)

plt.figure(figsize=(12, 3))

plt.subplot(1, 2, 1)
plt.plot(epoch_length, train_acc, label='Train Accuaray')
plt.plot(epoch_length, test_acc, label='Test Accuaray')
plt.legend(loc='lower right')
plt.title('Accurary')

plt.subplot(1, 2, 2)
plt.plot(epoch_length, train_loss, label='Train Loss')
plt.plot(epoch_length, test_loss, label='Test Loss')
plt.legend(loc='upper right')
plt.title('Loss')

plt.show()

​

​

测试集表现不是很理想，合理尝试变化不同的批次，会有不同效果

6、模型评估

# 评估：返回的是自己在model.compile中设置，这里为accuracy
test_acc, test_loss = test(test_dl, model, loss_fn)
print("socre[loss, accuracy]: ", test_acc, test_loss) # 返回为两个，一个是loss，一个是accuracy

socre[loss, accuracy]:  0.8064516129032258 0.5212066173553467