TypeError: __init__() missing 1 required positional argument: 'dtype'

My training command: python tools/train.py -f exps/example/yolox_voc/yolox_voc_s.py -d 1 -b 1 --fp16 -o -c weights/yolox_s.pth.tar
My training log:
2021-08-08 20:00:58.315 | INFO | yolox.core.trainer:before_epoch:193 - ---> start train epoch1
2021-08-08 20:00:58.315 | INFO | yolox.core.trainer:after_train:188 - Training of experiment is done and the best AP is 0.00
2021-08-08 20:00:58.316 | ERROR | yolox.core.launch:launch:90 - An error has been caught in function 'launch', process 'MainProcess' (9236), thread 'MainThread' (9960):
Traceback (most recent call last):

File "tools/train.py", line 125, in
args=(exp, args),
│ └ Namespace(batch_size=1, ckpt='weights/yolox_s.pth.tar', devices=1, dist_backend='nccl', dist_url=None, exp_file='exps/example...
└ ╒══════════════════╤═════════════════════════════════════════════════════════════════════════════════════════════════════════...

File "d:\mmdetection\yolox\yolox-main\yolox\core\launch.py", line 90, in launch
main_func(*args)
│ └ (╒══════════════════╤════════════════════════════════════════════════════════════════════════════════════════════════════════...
└ <function main at 0x000001E02979ED38>

File "tools/train.py", line 104, in main
trainer.train()
│ └ <function Trainer.train at 0x000001E03F0B9318>
└ <yolox.core.trainer.Trainer object at 0x000001E04024AF88>

File "d:\mmdetection\yolox\yolox-main\yolox\core\trainer.py", line 71, in train
self.train_in_epoch()
│ └ <function Trainer.train_in_epoch at 0x000001E03F3C90D8>
└ <yolox.core.trainer.Trainer object at 0x000001E04024AF88>

File "d:\mmdetection\yolox\yolox-main\yolox\core\trainer.py", line 80, in train_in_epoch
self.train_in_iter()
│ └ <function Trainer.train_in_iter at 0x000001E03F3C3288>
└ <yolox.core.trainer.Trainer object at 0x000001E04024AF88>

File "d:\mmdetection\yolox\yolox-main\yolox\core\trainer.py", line 86, in train_in_iter
self.train_one_iter()
│ └ <function Trainer.train_one_iter at 0x000001E0402123A8>
└ <yolox.core.trainer.Trainer object at 0x000001E04024AF88>

File "d:\mmdetection\yolox\yolox-main\yolox\core\trainer.py", line 92, in train_one_iter
inps, targets = self.prefetcher.next()
│ │ └ <function DataPrefetcher.next at 0x000001E03F3C3048>
│ └ <yolox.data.data_prefetcher.DataPrefetcher object at 0x000001E0497091C8>
└ <yolox.core.trainer.Trainer object at 0x000001E04024AF88>

File "d:\mmdetection\yolox\yolox-main\yolox\data\data_prefetcher.py", line 48, in next
self.preload()
│ └ <function DataPrefetcher.preload at 0x000001E03F3C1F78>
└ <yolox.data.data_prefetcher.DataPrefetcher object at 0x000001E0497091C8>

File "d:\mmdetection\yolox\yolox-main\yolox\data\data_prefetcher.py", line 30, in preload
self.next_input, self.next_target, _, _ = next(self.loader)
│ │ │ │ │ └ <torch.utils.data.dataloader._MultiProcessingDataLoaderIter object at 0x000001E0495AAE88>
│ │ │ │ └ <yolox.data.data_prefetcher.DataPrefetcher object at 0x000001E0497091C8>
│ │ │ └ tensor([[[ 11.0000, 529.3131, 509.4199, 221.3737, 261.1603],
│ │ │ [ 14.0000, 369.5336, 320.0000, 540.9327, 640.0000],
│ │ │ ...
│ │ └ <yolox.data.data_prefetcher.DataPrefetcher object at 0x000001E0497091C8>
│ └ tensor([[[[ 0.0878, 0.2258, 0.1085, ..., -0.3814, -0.3814, -0.3814],
│ [ 0.0257, 0.2189, 0.1568, ..., -0.3814, ...
└ <yolox.data.data_prefetcher.DataPrefetcher object at 0x000001E0497091C8>

File "D:\Anaconda\envs\yolox2\lib\site-packages\torch\utils\data\dataloader.py", line 435, in next
data = self._next_data()
│ └ <function _MultiProcessingDataLoaderIter._next_data at 0x000001E03EB90318>
└ <torch.utils.data.dataloader._MultiProcessingDataLoaderIter object at 0x000001E0495AAE88>

File "D:\Anaconda\envs\yolox2\lib\site-packages\torch\utils\data\dataloader.py", line 1085, in _next_data
return self._process_data(data)
│ │ └ <torch._utils.ExceptionWrapper object at 0x000001E04A1D66C8>
│ └ <function _MultiProcessingDataLoaderIter._process_data at 0x000001E03EB90438>
└ <torch.utils.data.dataloader._MultiProcessingDataLoaderIter object at 0x000001E0495AAE88>

File "D:\Anaconda\envs\yolox2\lib\site-packages\torch\utils\data\dataloader.py", line 1111, in _process_data
data.reraise()
│ └ <function ExceptionWrapper.reraise at 0x000001E02B690318>
└ <torch._utils.ExceptionWrapper object at 0x000001E04A1D66C8>

File "D:\Anaconda\envs\yolox2\lib\site-packages\torch_utils.py", line 428, in reraise
raise self.exc_type(msg)
│ │ └ 'Caught MemoryError in DataLoader worker process 1.\nOriginal Traceback (most recent call last):\n File "D:\Anaconda\envs...
│ └ <class 'numpy.core._exceptions._ArrayMemoryError'>
└ <torch._utils.ExceptionWrapper object at 0x000001E04A1D66C8>

TypeError: init() missing 1 required positional argument: 'dtype'
I don't know how to solve it

Start with the training command in tools/train.py and trace the failing data-loading path through yolox/core/trainer.py and yolox/data/data_prefetcher.py. Reproduce the failure, inspect the nested NumPy ArrayMemoryError in the DataLoader worker traceback, and consider the issue resolved when the command can load a batch and continue training without the reported exception.