深層学習メモ_3. 完全に接続されたニューラル ネットワークが MNIST データを解決する

1. パッケージをインポートする

import numpy as np
import torch
# 导入 pytorch 内置的 mnist 数据
from torchvision.datasets import mnist
# import torchvision
# 导入预处理模块
import torchvision.transforms as transforms
from torch.utils.data import DataLoader
# 导入nn及优化器
import torch.nn.functional as F
import torch.optim as optim
from torch import nn
# from torch.utils.tensorboard import SummaryWriter

2. いくつかのハイパーパラメータを定義し、PyTorchtorchvisionライブラリを使用して MNIST データ セットをダウンロードし、前処理します。

# 定义一些超参数
train_batch_size = 16
test_batch_size = 16
learning_rate = 0.01
num_epoches = 20

# 定义预处理函数
transform = transforms.Compose([transforms.ToTensor(),
                                transforms.Normalize([0.5],
                                                     [0.5])])
# 下载数据，并对数据进行预处理
train_dataset = mnist.MNIST('data',
                            train=True,
                            transform=transform,
                            download=True)
test_dataset = mnist.MNIST('data',
                           train=False,
                           transform=transform)
# 得到一个生成器
train_loader = DataLoader(train_dataset,
                          batch_size=train_batch_size,
                          shuffle=True)
test_loader = DataLoader(test_dataset,
                         batch_size=test_batch_size,
                         shuffle=False)

3. シーケンシャルを使用して完全に接続されたニューラル ネットワークを構築する

class Net(nn.Module):
    """
    使用sequential构建网络，Sequential()函数的功能是将网络的层组合到一起
    """

    def __init__(self, in_dim, n_hidden_1, n_hidden_2, out_dim):
        super(Net, self).__init__()
        self.flatten = nn.Flatten()
        self.layer1 = nn.Sequential(nn.Linear(in_dim, n_hidden_1), nn.BatchNorm1d(n_hidden_1))
        self.layer2 = nn.Sequential(nn.Linear(n_hidden_1, n_hidden_2), nn.BatchNorm1d(n_hidden_2))
        self.out = nn.Sequential(nn.Linear(n_hidden_2, out_dim))

    def forward(self, x):
        x = self.flatten(x)
        x = F.relu(self.layer1(x))
        x = F.relu(self.layer2(x))
        x = F.softmax(self.out(x), dim=1)
        return x

4. 損失関数とオプティマイザー: クロスエントロピー損失関数と確率的勾配降下 (SGD) オプティマイザー。

lr = 0.01
momentum = 0.9

# 实例化模型
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
model = Net(28 * 28, 300, 100, 10)
model.to(device)

# 定义损失函数和优化器
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(), lr=lr, momentum=momentum)

5. テスト セットでモデルのパフォーマンスをテストし、テストの損失と精度を記録します。

# 开始训练
losses = []
acces = []
eval_losses = []
eval_acces = []
for epoch in range(num_epoches):
    train_loss = 0
    train_acc = 0
    model.train()
    # 动态修改参数学习率
    if epoch % 5 == 0:
        optimizer.param_groups[0]['lr'] *= 0.9
        print("学习率:{:.6f}".format(optimizer.param_groups[0]['lr']))

    for img, label in train_loader:
        img = img.to(device)
        label = label.to(device)
        # 正向传播
        out = model(img)
        loss = criterion(out, label)
        # 反向传播
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
        # 记录误差
        train_loss += loss.item()
        # 计算分类的准确率
        _, pred = out.max(1)
        num_correct = (pred == label).sum().item()
        acc = num_correct / img.shape[0]
        train_acc += acc

    losses.append(train_loss / len(train_loader))
    acces.append(train_acc / len(train_loader))

    # 在测试集上检验效果
    eval_loss = 0
    eval_acc = 0
    # net.eval() # 将模型改为预测模式
    model.eval()
    for img, label in test_loader:
        img = img.to(device)
        label = label.to(device)
        img = img.view(img.size(0), -1)
        out = model(img)
        loss = criterion(out, label)
        # 记录误差
        eval_loss += loss.item()
        # 记录准确率
        _, pred = out.max(1)
        num_correct = (pred == label).sum().item()
        acc = num_correct / img.shape[0]
        eval_acc += acc

    eval_losses.append(eval_loss / len(test_loader))
    eval_acces.append(eval_acc / len(test_loader))
    print('epoch: {}, Train Loss: {:.4f}, Train Acc: {:.4f}, Test Loss: {:.4f}, Test Acc: {:.4f}'
          .format(epoch, train_loss / len(train_loader), train_acc / len(train_loader),
                  eval_loss / len(test_loader), eval_acc / len(test_loader)))
    

 

6. loss、access、eval_losses、および eval_acces を CSV ファイルに保存する

import csv
# 定义CSV文件的路径
csv_file = 'metrics.csv'
# 将数据保存到CSV文件
with open(csv_file, 'w', newline='') as file:
    writer = csv.writer(file)
    # 写入列名
    writer.writerow(["Epoch", "Loss", "Accuracy", "Eval Loss", "Eval Accuracy"])
    # 写入数据行
    for epoch in range(len(losses)):
        loss = losses[epoch]
        acc = acces[epoch]
        eval_loss = eval_losses[epoch]
        eval_acc = eval_acces[epoch]
        writer.writerow([epoch+1, loss, acc, eval_loss, eval_acc])

7. 損失画像と ACC 画像の描画

import pandas as pd
import matplotlib.pyplot as plt

# 读取文件
df = pd.read_csv('metrics.csv')

# 提取数据
epochs = df['Epoch']
losses = df['Loss']
accuracies = df['Accuracy']
eval_losses = df['Eval Loss']
eval_accuracies = df['Eval Accuracy']

# 绘制曲线图
plt.figure(figsize=(10, 6))
plt.plot(epochs, losses, label='Loss')
plt.plot(epochs, eval_losses, label='Eval Loss')
plt.xlabel('Epoch')
plt.ylabel('Value')
plt.title('Loss and Accuracy')
plt.legend()
plt.show() 


# 绘制曲线图
plt.figure(figsize=(10, 6))
plt.plot(epochs, accuracies, label='Accuracy')
plt.plot(epochs, eval_accuracies, label='Eval Accuracy')
plt.xlabel('Epoch')
plt.ylabel('Value')
plt.title('Loss and Accuracy')
plt.legend()
plt.show()