Invalid label type

[ghost]

Hi @chrischoy,

I was training on a subset of scannet & I got the following error;

Traceback (most recent call last):
  File "main.py", line 154, in <module>
    main()
  File "main.py", line 147, in main
    train(model, train_data_loader, val_data_loader, config)
  File "/home/x23/workspace_pcs/SpatioTemporalSegmentation/lib/train.py", line 78, in train
    coords, input, target = data_iter.next()
  File "/home/x23/miniconda3/envs/mink/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 819, in __next__
    return self._process_data(data)
  File "/home/x23/miniconda3/envs/mink/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 846, in _process_data
    data.reraise()
  File "/home/x23/miniconda3/envs/mink/lib/python3.7/site-packages/torch/_utils.py", line 369, in reraise
    raise self.exc_type(msg)
AssertionError: Caught AssertionError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "/home/x23/miniconda3/envs/mink/lib/python3.7/site-packages/torch/utils/data/_utils/worker.py", line 178, in _worker_loop
    data = fetcher.fetch(index)
  File "/home/x23/miniconda3/envs/mink/lib/python3.7/site-packages/torch/utils/data/_utils/fetch.py", line 44, in fetch
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/home/x23/miniconda3/envs/mink/lib/python3.7/site-packages/torch/utils/data/_utils/fetch.py", line 44, in <listcomp>
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "/home/x23/workspace_pcs/SpatioTemporalSegmentation/lib/dataset.py", line 266, in __getitem__
    coords, feats, labels, center=center)
  File "/home/x23/workspace_pcs/SpatioTemporalSegmentation/lib/voxelizer.py", line 130, in voxelize
    coords_aug, feats, labels = ME.utils.sparse_quantize(coords_aug, feats, labels=labels, ignore_label=self.ignore_label)
  File "/home/x23/miniconda3/envs/mink/lib/python3.7/site-packages/MinkowskiEngine-0.3.3-py3.7-linux-x86_64.egg/MinkowskiEngine/utils/quantization.py", line 194, in sparse_quantize
    ignore_label)
  File "/home/x23/miniconda3/envs/mink/lib/python3.7/site-packages/MinkowskiEngine-0.3.3-py3.7-linux-x86_64.egg/MinkowskiEngine/utils/quantization.py", line 82, in quantize_label
    assert labels.dtype == np.int32, f"Invalid label type {labels.dtype} != np.int32"
AssertionError: Invalid label type float32 != np.int32

[ghost]

As a workaround, I changed the dtype of labels to int32 before feeding them to ME.utils.sparse_quantize, see below;

labels = labels.astype(np.int32, copy=False)
coords_aug, feats, labels = ME.utils.sparse_quantize(coords_aug, feats, labels=labels, ignore_label=self.ignore_label)