如何在图像分类pytorch中提前停止

import torch
import torch.nn as nn
import torch.optim as optim
from torchvision import datasets, transforms

# 定义模型
model = ...

# 定义损失函数和优化器
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(), lr=0.01)

# 定义训练和验证函数
def train(model, train_loader):
    ...

def validate(model, val_loader):
    ...

# 定义提前停止的函数
def early_stopping(val_acc_history, threshold, patience):
    if len(val_acc_history) < patience:
        return False
    for i in range(1, patience+1):
        if val_acc_history[-i] > threshold:
            return False
    return True

# 加载数据集
train_dataset = datasets.ImageFolder('train_dir', transform=transforms.ToTensor())
val_dataset = datasets.ImageFolder('val_dir', transform=transforms.ToTensor())
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=64, shuffle=True)
val_loader = torch.utils.data.DataLoader(val_dataset, batch_size=64, shuffle=False)

# 训练过程
max_epochs = 100
threshold = 0.95
patience = 5
best_val_acc = 0.0
val_acc_history = []

for epoch in range(max_epochs):
    train(model, train_loader)
    val_acc = validate(model, val_loader)
    val_acc_history.append(val_acc)
    
    if val_acc > best_val_acc:
        best_val_acc = val_acc
        # 保存最佳模型
        
    if early_stopping(val_acc_history, threshold, patience):
        print("Early stopping triggered.")
        break