AttributeError：“_MultiProcessingDataLoaderIter”对象没有属性“next”

Question

我尝试使用

Torch Dataset and DataLoader

加载数据集，但出现以下错误：

AttributeError: '_MultiProcessingDataLoaderIter' object has no attribute 'next'

我使用的代码是：

class WineDataset(Dataset):

    def __init__(self):
        # Initialize data, download, etc.
        # read with numpy or pandas
        xy = np.loadtxt('./data/wine.csv', delimiter=',', dtype=np.float32, skiprows=1)
        self.n_samples = xy.shape[0]

        # here the first column is the class label, the rest are the features
        self.x_data = torch.from_numpy(xy[:, 1:]) # size [n_samples, n_features]
        self.y_data = torch.from_numpy(xy[:, [0]]) # size [n_samples, 1]

    # support indexing such that dataset[i] can be used to get i-th sample
    def __getitem__(self, index):
        return self.x_data[index], self.y_data[index]

    # we can call len(dataset) to return the size
    def __len__(self):
        return self.n_samples

    dataset = WineDataset()
        
    train_loader = DataLoader(dataset=dataset,
                              batch_size=4,
                              shuffle=True,
                              num_workers=2)

我尝试让num_workers=0，仍然有同样的错误。

Python version 3.8.9
PyTorch version 1.13.0

Answer 1

当我尝试调用 next() 方法时，我也遇到了同样的问题

dataiter = iter(dataloader)
data = dataiter.next()

您需要使用以下内容，它可以完美运行：

dataiter = iter(dataloader)
data = next(dataiter)

最后你的代码应该如下所示：

class WineDataset(Dataset):

    def __init__(self):
        # Initialize data, download, etc.
        # read with numpy or pandas
        xy = np.loadtxt('./data/wine.csv', delimiter=',', dtype=np.float32, skiprows=1)
        self.n_samples = xy.shape[0]

        # here the first column is the class label, the rest are the features
        self.x_data = torch.from_numpy(xy[:, 1:]) # size [n_samples, n_features]
        self.y_data = torch.from_numpy(xy[:, [0]]) # size [n_samples, 1]

    # support indexing such that dataset[i] can be used to get i-th sample
    def __getitem__(self, index):
        return self.x_data[index], self.y_data[index]

    # we can call len(dataset) to return the size
    def __len__(self):
        return self.n_samples

    dataset = WineDataset()
        
    train_loader = DataLoader(dataset=dataset,
                              batch_size=4,
                              shuffle=True,
                              num_workers=2)

dataiter = iter(dataloader)
data = next(dataiter)

Answer 2

在 pytorch 1.12 中语法：

iter(trn_loader).next()

工作正常以及：

next(iter(trn_loader))

从 pytorch 1.13 开始，唯一有效的语法是：

next(iter(trn_loader))

Answer 3

于 2023 年 4 月更新 而不是从

iter(trn_loader).next()

更改 到

next(iter(trn_loader))

。我更喜欢解决 pyTorch 版本问题，因为我不知道代码中有多少个

.next()

。

conda install pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 -c pytorch

Answer 4

我也面临着同样的问题。我使用的是torch ==2.3版本。 我收到的错误是 RuntimeError: An attempts has made to start a new process before the current process has finish its bootstrapping stage，这表明 Python 中的多处理存在问题。当使用 PyTorch 的 DataLoader 启用多处理时，经常会发生此错误，尤其是在 Windows 系统上。 这对我有用：

类 WineDataset（数据集）：

def __init__(self):                
    # data loading
    xy = np.loadtxt('E:\\Github\\Pytorch_beginner\\wine.csv', delimiter = ",", dtype = np.float32, skiprows =1 )
    self.x = torch.from_numpy(xy[:,1:])
    self.y = torch.from_numpy(xy[:,[0]]) # n_samples,1
    self.n_samples = xy.shape[0]

# support indexing such that dataset[i] can be used to get i-th sample
def __getitem__(self,index):
    # dataset[0]
    return self.x[index], self.y[index]

# we can call len(dataset) to return the size     
def __len__(self):
    #len(dataset)
    return self.n_samples  

if name == 'main': 数据集 = WineDataset() dataloader = DataLoader（数据集=数据集，batch_size=4，shuffle=True，num_workers=2） dataiter = iter(数据加载器) 数据 = 下一个（dataiter）

希望这有帮助！