batch_size = coors[-1, 0].item() + 1 IndexError: index -1 is out of bounds for dimension 0 with size 0 pointpillars

### Checklist

- [X] I have searched for [similar issues](https://github.com/isl-org/Open3D-ML/issues).
- [X] I have tested with the [latest development wheel](http://www.open3d.org/docs/latest/getting_started.html#development-version-pip).
- [X] I have checked the [release documentation](http://www.open3d.org/docs/release/) and the [latest documentation](http://www.open3d.org/docs/latest/) (for `master` branch).

### My Question

points: [tensor([[-54.6733, -35.9321, 65.5350, 0.5000],
[-54.5223, -35.9321, 65.5350, 0.5000],
[-54.3713, -35.9321, 65.5350, 0.5000],
...,
[ 3.0787, 1.8103, 3.2590, 0.5000],
[ 3.0786, 1.8058, 3.2510, 0.5000],
[ 3.0785, 1.8014, 3.2430, 0.5000]], device='cuda:0')]
voxels: tensor([], device='cuda:0', size=(0, 32, 4))
numpts: tensor([], device='cuda:0', dtype=torch.int64)
coors: tensor([], device='cuda:0', size=(0, 4), dtype=torch.int32)
50%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████ | 1/2 [00:01<00:01, 1.05s/it]
Traceback (most recent call last):
File "detection_torch.py", line 138, in
result = pipeline.run_inference(data)[0]
File "/home/lnt/miniconda3/envs/open3d/lib/python3.7/site-packages/open3d/_ml3d/torch/pipelines/object_detection.py", line 72, in run_inference
results = model(data)
File "/home/lnt/miniconda3/envs/open3d/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl
return forward_call(*input, **kwargs)
File "/home/lnt/miniconda3/envs/open3d/lib/python3.7/site-packages/open3d/_ml3d/torch/models/point_pillars.py", line 137, in forward
x = self.extract_feats(inputs)
File "/home/lnt/miniconda3/envs/open3d/lib/python3.7/site-packages/open3d/_ml3d/torch/models/point_pillars.py", line 111, in extract_feats
**batch_size = coors[-1, 0].item() + 1
IndexError: index -1 is out of bounds for dimension 0 with size 0**

Struggling to get this working with my own pcd file data (RealSense D455)
I tried changing the voxel range but that didn't work either. I can't seem to get voxel or coors results.

Seems to work with other example pcd data.