RuntimeError: Could not infer dtype of numpy.uint8

[FreshmanCow]

在训练自定义的数据集的时候，为什么要出现以下错误呢，是python版本不对吗

Traceback (most recent call last): File "/media/s304/Data/jjw/RT-DETR-main/rtdetr_pytorch/tools/train.py", line 49, in main(args) File "/media/s304/Data/jjw/RT-DETR-main/rtdetr_pytorch/tools/train.py", line 36, in main solver.fit() File "/media/s304/Data/jjw/RT-DETR-main/rtdetr_pytorch/tools/../src/solver/det_solver.py", line 37, in fit train_stats = train_one_epoch( File "/media/s304/Data/jjw/RT-DETR-main/rtdetr_pytorch/tools/../src/solver/det_engine.py", line 35, in train_one_epoch for samples, targets in metric_logger.log_every(data_loader, print_freq, header): File "/media/s304/Data/jjw/RT-DETR-main/rtdetr_pytorch/tools/../src/misc/logger.py", line 215, in log_every for obj in iterable: File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 633, in next data = self._next_data() File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 677, in _next_data data = self._dataset_fetcher.fetch(index) # may raise StopIteration File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 51, in fetch data = [self.dataset[idx] for idx in possibly_batched_index] File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 51, in data = [self.dataset[idx] for idx in possibly_batched_index] File "/media/s304/Data/jjw/RT-DETR-main/rtdetr_pytorch/tools/../src/data/coco/coco_dataset.py", line 54, in getitem img, target = self._transforms(img, target) File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(*args, *kwargs) File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torchvision/transforms/v2/_container.py", line 51, in forward sample = transform(sample) File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(args, **kwargs) File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torchvision/transforms/v2/_transform.py", line 44, in forward flat_outputs = [ File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torchvision/transforms/v2/_transform.py", line 45, in self._transform(inpt, params) if needs_transform else inpt File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torchvision/transforms/v2/_color.py", line 273, in _transform inpt = self._permute_channels(inpt, permutation=params["channel_permutation"]) File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torchvision/transforms/v2/_color.py", line 249, in _permute_channels inpt = F.pil_to_tensor(inpt) File "/media/s304/Data/anaconda3/envs/detr/lib/python3.9/site-packages/torchvision/transforms/functional.py", line 207, in pil_to_tensor img = torch.as_tensor(np.array(pic, copy=True)) RuntimeError: Could not infer dtype of numpy.uint8

[beyond-ssf]

我也是训练自己的单类别数据集在同样的文件位置报了同样的错误，请问你是怎么解决的呢？

[adensur]

For me, installing numpy 1.* version solved the issue. Similar discussions: https://github.com/facebookresearch/segment-anything/issues/477