Why can't iter dataloader?

data
1

when i use the function: prefetch, i found the the code runtime error
The code like this:
class data_prefetch():
def init(self, cfg, loader, is_train):
self.loader = loader

def preload(self):
    try:
        self.next_meta = next(self.loader)
   except:
        self.next_meta = None
        return

the function run and come to the except.

or like this:
class data_prefetch():
def init(self, cfg, loader, is_train):
self.loader = iter(loader)

def preload(self):
    try:
        self.next_meta = next(self.loader)

when i used the second class, the object come to runtime error?

why and how to solve this problem?

2

import torchvision.transforms as transforms
class data_prefetch():
def init(self, cfg, loader, is_train):
self.loader = loader
self.is_train = is_train
self.preload()

def preload(self):
    try:
        self.next_meta = next(self.loader)
        print("try is okey")
    except:
        self.next_meta = None

def next(self):
    meta = self.next_meta
    self.preload()
    return meta

def __len__(self):
    return len(self.loader)
3

Calling next on an iterator is expected to raise a StopIteration which you would need to handle as seen here:

dataset = TensorDataset(torch.randn(5))
loader = DataLoader(dataset)
iter_loader = iter(loader)

for _ in range(6):
    x = next(iter_loader)
# StopIteration
4

Thanks，maybe other question about custum dataset or dataloader.