Is there a way to get item with AspectRatioGroupedDataset?

[Keiku]

The following is an example. I think you can get item with a normal data loader, but what is the correct way to do it with detectron2?

---------------------------------------------------------------------------
NotImplementedError                       Traceback (most recent call last)
/tmp/ipykernel_7533/1118859689.py in <module>
     31 
     32     train_data_loader = build_detection_train_loader(cfg, mapper)
---> 33     print(train_data_loader[0])
     34     i = 0
     35     for batch in train_data_loader:

~/.pyenv/versions/3.8.6-mask-former/lib/python3.8/site-packages/torch/utils/data/dataset.py in __getitem__(self, index)
     31 
     32     def __getitem__(self, index) -> T_co:
---> 33         raise NotImplementedError
     34 
     35     def __add__(self, other: 'Dataset[T_co]') -> 'ConcatDataset[T_co]':

NotImplementedError: 

[ppwwyyxx]

There is no way, and same for "normal data loader" as well:

$cat a.py
─────┬────────────────────────────────────────────────────────────────────────────────────
   1 │
   2 │ import torch
   3 │ from detectron2.data import DatasetFromList
   4 │ d = DatasetFromList([1, 2, 3])
   5 │ d = torch.utils.data.DataLoader(d, batch_size=2)
   6 │ print(d[1])
─────┴────────────────────────────────────────────────────────────────────────────────────
$python3 a.py
Traceback (most recent call last):
  File "a.py", line 6, in <module>
    print(d[1])
TypeError: 'DataLoader' object is not subscriptable