RuntimeError: min(): Expected reduction dim to be specified for input.numel() == 0.

[SenHaiFeiXia]

torch.Size([72902, 3]) [04/07 20:49:20] selected_pts_mask [04/07 20:49:20] tensor([False, False, False, ..., False, False, False], device='cuda:0') [04/07 20:49:20] selected_pts_mask.shape [04/07 20:49:20] torch.Size([72902]) [04/07 20:49:20] stds [04/07 20:49:20] tensor([], device='cuda:0', size=(0, 3)) [04/07 20:49:20] stds.shape [04/07 20:49:20] torch.Size([0, 3]) [04/07 20:49:20] means [04/07 20:49:20] tensor([], device='cuda:0', size=(0, 3)) [04/07 20:49:20] means.shape [04/07 20:49:20] torch.Size([0, 3]) [04/07 20:49:20] Traceback (most recent call last): File "train.py", line 219, in training(lp.extract(args), op.extract(args), pp.extract(args), args.test_iterations, args.save_iterations, args.checkpoint_iterations, args.start_checkpoint, args.debug_from) File "train.py", line 120, in training gaussians.densify_and_prune(opt.densify_grad_threshold, 0.005, scene.cameras_extent, size_threshold) File "/home/hdd1/gaussian-splatting/scene/gaussian_model.py", line 424, in densify_and_prune self.densify_and_split(grads, max_grad, extent) File "/home/hdd1/gaussian-splatting/scene/gaussian_model.py", line 390, in densify_and_split samples = torch.normal(mean=means, std=stds) RuntimeError: min(): Expected reduction dim to be specified for input.numel() == 0. Specify the reduction dim with the 'dim' argument. Training progress: 29%|█████████████████████████████████████████████████████████▏ | 8800/30000 [01:11<02:52, 123.04it/s, Loss=0.0022076] Probablims during densify_and_split ,There is no Gaussian that can be used for splitting， Does this mean overfitting?