Pytorch-报错：RuntimeError: expected scalar type Long but found Int

报错代码

def train(n_epochs, model, train_x, train_y, test_x, max_cycle_t, y_test):
    rmse_history = []
    for epoch in range(1, n_epochs + 1):
        model.train()

        x_train, y_train = Variable(train_x), Variable(train_y)

        optimizer.zero_grad()

        output_train = model(x_train)

        loss_train = criterion(output_train, y_train)
        loss_train.backward()
        optimizer.step()

应该是y_train标签类型有错误，找到dataset_prepare文件，在后面加上.long()

    train_y = train_y.astype(int)
    train_y = torch.from_numpy(train_y).long()

问题解决。