加载ResNet18的.pth模型文件时参数形状不匹配错误的解决求助

各位好，我最近遇到了一个模型加载的棘手问题，想请教一下大家的解决方案：

我训练了一个ResNet18模型并保存成了.pth文件，但尝试加载的时候一直报参数形状不匹配的错误，错误信息如下，后面还有好几行类似的格式：

Error loading checkpoint: Error(s) in loading state_dict for ResNet:
size mismatch for layer1.0.conv1.weight: copying a param with shape torch.Size([64, 64, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for layer1.1.conv1.weight: copying a param with shape torch.Size([64, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 64, 3, 3]).
size mismatch for layer2.0.conv1.weight: copying a param with shape torch.Size([128, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([128, 64, 3, 3]).
size mismatch for layer2.0.downsample.0.weight: copying a param with shape torch.Size([512, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([128, 64, 1, 1]).
size mismatch for layer2.0.downsample.1.weight: copying a param with shape torch.Size([512]) from checkpoint, the shape in current model is torch.Size([128]).

这是我当初训练模型的代码：

teacher = models.resnet18(pretrained=True)

num_features = teacher.fc.in_features
teacher.fc = nn.Linear(num_features, 5)


criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(teacher.parameters(), lr=0.0001, momentum=0.9, weight_decay=0.0001)

scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=7, gamma=0.1)

def train_and_evaluate(model, train_loader, val_loader, criterion, optimizer, num_epochs, lambda_l1, learning_rate):
    for epoch in range(num_epochs):
        model.train()
        for images, labels in train_loader:
            # Filter out class 2 samples during training
            mask = labels != 2
            images, labels = images[mask], labels[mask]

            if len(labels) == 0: 
                continue

            optimizer.zero_grad()
            outputs = model(images)
            loss = criterion(outputs, labels)
            total_loss = loss 
            total_loss.backward()
            optimizer.step()

        # Evaluate on both training and validation sets, excluding class 2
        train_loss, train_accuracy = evaluate_model(model, train_loader, criterion)
        val_loss, val_accuracy = evaluate_model(model, val_loader, criterion)

        print(f"Epoch {epoch+1} - Training Loss: {train_loss:.4f}, Training Accuracy: {train_accuracy:.4%}, Validation Loss: {val_loss:.4f}, Validation Accuracy: {val_accuracy:.4%}")

我知道训练时排除第二类的逻辑写得有点乱，但这个模型训练花了超级久，我真的很想把它救回来😭。

另外，我加载模型的代码是这样的：

checkpoint = torch.load("model968acc.pth", map_location="cpu")
teacher.load_state_dict(checkpoint, strict=False)

如果这个模型真的没法抢救了，也想请大家推荐一些靠谱的模型保存方法，避免以后再遇到这种加载失败的问题！

备注：内容来源于stack exchange，提问作者Xayyna