【error】raise NotImplementedError

错误：

Traceback (most recent call last):。
  File "main.py", line 103, in train
    for batch_i, (images, labels) in enumerate(train_loader, start=1):
  File "/u。。/anaconda2/lib/python2.7/site-packages/torch/utils/data/dataloader.py", line 310, in __iter__
    return DataLoaderIter(self)
  File "/us。。/anaconda2/lib/python2.7/site-packages/torch/utils/data/dataloader.py", line 180, in __init__
    self._put_indices()
  File "/us。。/anaconda2/lib/python2.7/site-packages/torch/utils/data/dataloader.py", line 219, in _put_indices
    indices = next(self.sample_iter, None)
  File "/u。。/anaconda2/lib/python2.7/site-packages/torch/utils/data/sampler.py", line 119, in __iter__
    for idx in self.sampler:
  File "/u。。/anaconda2/lib/python2.7/site-packages/torch/utils/data/sampler.py", line 50, in __iter__
    return iter(torch.randperm(len(self.data_source)).long())
  File "/us。。/anaconda2/lib/python2.7/site-packages/torch/utils/data/dataset.py", line 16, in __len__
    raise NotImplementedError
NotImplementedError

未实现的错误

再看train_loader部分：

代码如下

train_loader = torch.utils.data.DataLoader(dataset=coco_train,  batch_size=opt.batch_size,
                                              shuffle=shuffle, num_workers=num_workers, collate_fn=collate_fn)

coco_train = CocoDataset(opt, vocab=vocab, imagef='resizedtrain2014', 
                            annotf='coco/annotations/captions_train2014.json', transform=transform)

CocoDataset类定义如下：

class CocoDataset(data.Dataset):
    """COCO Custom Dataset compatible with torch.utils.data.DataLoader."""    
    def __init__(self, opt, vocab, imagef, annotf=None, transform=None):  
        。。。
    def __getitem__(self, index):
        """Returns one data pair (image and caption).""" 
        。。。
        return image, img_id

补上：

    def __len__(self):
        return len(self.。。。) 

重新运行，错误消失

原因是torch.utils.data.DataLoader中传入的参数dataset有要求，需要def __len__(self)函数。