getting segmentation fault (core dumped) at stage 1

[Matiur Rahman Minar]

Thanks for the outstanding project. Can you please help me regarding this issue? I am getting segmentation fault every time training in the stage 1. Stage 0 seems to be okay. Thanks. Here's the error trace below:

[INFO] Loading stage 0 model to init stage 1 ...

[INFO] loaded model.

[WARN] missing keys: ['vertices_offsets']

[INFO] Loading latest checkpoint ...

[WARN] No checkpoint found, abort loading latest model.

[INFO] ColmapDataset: image H = 1016, W = 1547

[INFO] 93 image exists in all 93 colmap entries.

[INFO] use mask under folder data/custom/mask

[INFO] ColmapDataset: load poses (93, 4, 4), points (45633, 3)

[INFO] ColmapDataset: auto-scale 0.3562

[WARN] ColmapDataset: estimated AABB [-33.127134724523515, -22.970199505940037, -3.06402499043361, 37.161698704312926, 23.72691525611831, 14.77761798452963] exceeds provided bound 1.0! Consider improving --
bound to make scene included in trainable region.

[INFO] extracting sparse depth info...
100%|███████████████████████████████████████████████████████████████| 93/93 [00:00<00:00, 4141.71it/s]

[INFO] extracted 2716.57 valid sparse depth on average per image

Loading val data: 100%|███████████████████████████████████████████████| 12/12 [00:00<00:00, 52.23it/s]

[mark untrained grid] 0 from 2097152

==> Start Training Epoch 1, lr=0.000100 ...
  0% 0/81 [00:00<?, ?it/s]Segmentation fault (core dumped)

[Matiur Rahman Minar]

Fixed the issue by installing the torch and torch-scatter with the same CUDA version. Thanks.