ValueError: some of the strides of a given numpy array are negative. This is currently not supported, but will be added in future releases.

[zhixuanli]

@YutingXiao

I'm appreciated for your impressive work!

After configuring everything as your instructions, I get the following error:

[01/10 16:58:55 d2.engine.train_loop]: Starting training from iteration 0
[01/10 16:58:56 d2.engine.hooks]: Total training time: 0:00:00 (0:00:00 on hooks)
Traceback (most recent call last):
  File "tools/train_net.py", line 173, in <module>
    args=(args,),
  File "/home/lzx/Amodal-Segmentation-Based-on-Visible-Region-Segmentation-and-Shape-Prior-main/detectron2/engine/launch.py", line 51, in launch
    main_func(*args)
  File "tools/train_net.py", line 161, in main
    return trainer.train()
  File "/home/lzx/Amodal-Segmentation-Based-on-Visible-Region-Segmentation-and-Shape-Prior-main/detectron2/engine/defaults.py", line 416, in train
    super().train(self.start_iter, self.max_iter)
  File "/home/lzx/Amodal-Segmentation-Based-on-Visible-Region-Segmentation-and-Shape-Prior-main/detectron2/engine/train_loop.py", line 132, in train
    self.run_step()
  File "/home/lzx/Amodal-Segmentation-Based-on-Visible-Region-Segmentation-and-Shape-Prior-main/detectron2/engine/train_loop.py", line 211, in run_step
    data = next(self._data_loader_iter)
  File "/home/lzx/anaconda3/envs/aaai21amodal/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 345, in __next__
    data = self._next_data()
  File "/home/lzx/anaconda3/envs/aaai21amodal/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 838, in _next_data
    return self._process_data(data)
  File "/home/lzx/anaconda3/envs/aaai21amodal/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 881, in _process_data
    data.reraise()
  File "/home/lzx/anaconda3/envs/aaai21amodal/lib/python3.6/site-packages/torch/_utils.py", line 394, in reraise
    raise self.exc_type(msg)
ValueError: Caught ValueError in DataLoader worker process 1.
Original Traceback (most recent call last):
  File "/home/lzx/anaconda3/envs/aaai21amodal/lib/python3.6/site-packages/torch/utils/data/_utils/worker.py", line 178, in _worker_loop
    data = fetcher.fetch(index)
  File "/home/lzx/anaconda3/envs/aaai21amodal/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py", line 44, in fetch
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/home/lzx/anaconda3/envs/aaai21amodal/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py", line 44, in <listcomp>
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/home/lzx/Amodal-Segmentation-Based-on-Visible-Region-Segmentation-and-Shape-Prior-main/detectron2/data/common.py", line 39, in __getitem__
    data = self._map_func(self._dataset[cur_idx])
  File "/home/lzx/Amodal-Segmentation-Based-on-Visible-Region-Segmentation-and-Shape-Prior-main/detectron2/utils/serialize.py", line 23, in __call__
    return self._obj(*args, **kwargs)
  File "/home/lzx/Amodal-Segmentation-Based-on-Visible-Region-Segmentation-and-Shape-Prior-main/detectron2/data/dataset_mapper.py", line 131, in __call__
    annos, image_shape, mask_format=self.mask_format
  File "/home/lzx/Amodal-Segmentation-Based-on-Visible-Region-Segmentation-and-Shape-Prior-main/detectron2/data/detection_utils.py", line 284, in annotations_to_instances
    visible_masks = BitMasks(torch.stack([torch.from_numpy(x) for x in visible_masks]))
  File "/home/lzx/Amodal-Segmentation-Based-on-Visible-Region-Segmentation-and-Shape-Prior-main/detectron2/data/detection_utils.py", line 284, in <listcomp>
    visible_masks = BitMasks(torch.stack([torch.from_numpy(x) for x in visible_masks]))
ValueError: some of the strides of a given numpy array are negative. This is currently not supported, but will be added in future releases.

I've done some searches online but still get stuck.

Reference Link: https://discuss.pytorch.org/t/torch-from-numpy-not-support-negative-strides/3663

Can you give me some advice? Thanks a lot! ^_^

[YutingXiao]

Thanks!
I suggest you to check the version of torch and numpy, especially the torch. Is it 1.4??

[YutingXiao]

By the way, the torch package contains its own cuda, you can use torch.version.cuda to check whether it is the same as your own cuda vesion.

I remember that i met this error before. The reason is that i used wrong version of some packages.

[YutingXiao]

I think may be the reason is that when you use pip install -r requirement.txt, you download torch==1.4.0 based on cuda 10.1.

But the cuda version in your computer maybe is 11.0 or 10.0??

[zhixuanli]

@YutingXiao

Thank you for your replying.

After add .copy , like:

visible_masks = BitMasks(torch.stack([torch.from_numpy(x.copy() for x in visible_masks]))

This error is gone now!

I'm running the training script now, and it's easy for deployment.

Again, thank you for your open-source code !