RuntimeError: Trying to resize storage that is not resizable

Traceback (most recent call last):
  File "/home/wayshow/Multimodel/train/train2.py", line 243, in <module>
    main(args=parser.parse_args())
  File "/home/wayshow/Multimodel/train/train2.py", line 219, in main
    loss_t= train_step(args, model, train_loader, optimizer, scheduler, criterion, scaler, epoch)
  File "/home/wayshow/Multimodel/train/train2.py", line 85, in train_step
    for step, batch in enumerate(train_loader):
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 633, in __next__
    data = self._next_data()
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 1345, in _next_data
    return self._process_data(data)
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/dataloader.py", line 1371, in _process_data
    data.reraise()
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/_utils.py", line 644, in reraise
    raise exception
RuntimeError: Caught RuntimeError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/_utils/worker.py", line 308, in _worker_loop
    data = fetcher.fetch(index)
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/_utils/fetch.py", line 54, in fetch
    return self.collate_fn(data)
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/_utils/collate.py", line 265, in default_collate
    return collate(batch, collate_fn_map=default_collate_fn_map)
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/_utils/collate.py", line 142, in collate
    return [collate(samples, collate_fn_map=collate_fn_map) for samples in transposed]  # Backwards compatibility.
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/_utils/collate.py", line 142, in <listcomp>
    return [collate(samples, collate_fn_map=collate_fn_map) for samples in transposed]  # Backwards compatibility.
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/_utils/collate.py", line 145, in collate
    return elem_type([collate(samples, collate_fn_map=collate_fn_map) for samples in transposed])
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/_utils/collate.py", line 145, in <listcomp>
    return elem_type([collate(samples, collate_fn_map=collate_fn_map) for samples in transposed])
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/_utils/collate.py", line 123, in collate
    return collate_fn_map[collate_type](batch, collate_fn_map=collate_fn_map)
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/utils/data/_utils/collate.py", line 161, in collate_tensor_fn
    out = elem.new(storage).resize_(len(batch), *list(elem.size()))
  File "/home/wayshow/.conda/envs/cws/lib/python3.10/site-packages/monai/data/meta_tensor.py", line 268, in __torch_function__
    ret = super().__torch_function__(func, types, args, kwargs)
  File "/home/wayshow/.local/lib/python3.10/site-packages/torch/_tensor.py", line 1295, in __torch_function__
    ret = func(*args, **kwargs)
RuntimeError: Trying to resize storage that is not resizable

今天模型训练，遇到了个bug，，错误信息显示跟数据集的尺寸size有关，这位博主提到了数据维度不统一的问题，于是我在dataset类的__getitem__函数中对每个经过transform后的图像进行了输出，发现因为原图尺寸问题，裁剪为（96，96，48）的图像时，因为图像本身小于这个尺寸（裁剪超出边界）导致输出结果不是（96，96，48）