Error: cannot reshape tensor of 0 elements into shape [0, -1] because the unspecified dimension size -1 can be any value and is ambiguous

[evonneng]

new error with instant_ngp whenever viewer is enabled (seems like it's only on the viewer side?)

[tancik]

Hmm, Im not seeing the error. It is possible that the occupancy grid hasn't been initialized by the callback if the viewer is called before the first training step is taken.

[evonneng]

I see! :O i can take a look into this... But this happens constantly after the viewer is initialized and is well into training. It doesnt break the viewer but it just constantly occurs

[evonneng]

seems to be fixed

[kevinchan04]

Hi, the problem it seems still exist. Do you have any idea to debug it? Thanks!

[kevinchan04]

Sorry that I was not plasted the error message for more explanation. The detail of error message is:

Traceback (most recent call last): File "/opt/conda/bin/ns-train", line 8, in sys.exit(entrypoint()) File "/workspace/nerfstudio/nerfstudio/scripts/train.py", line 247, in entrypoint main( File "/workspace/nerfstudio/nerfstudio/scripts/train.py", line 233, in main launch( File "/workspace/nerfstudio/nerfstudio/scripts/train.py", line 172, in launch main_func(local_rank=0, world_size=world_size, config=config) File "/workspace/nerfstudio/nerfstudio/scripts/train.py", line 87, in train_loop trainer.train() File "/workspace/nerfstudio/nerfstudio/nerfstudio/engine/trainer.py", line 271, in train self.eval_iteration(step) File "/workspace/nerfstudio/nerfstudio/nerfstudio/utils/decorators.py", line 70, in wrapper ret = func(self, *args, kwargs) File "/workspace/nerfstudio/nerfstudio/nerfstudio/utils/profiler.py", line 93, in inner out = func(*args, *kwargs) File "/workspace/nerfstudio/nerfstudio/nerfstudio/engine/trainer.py", line 463, in eval_iteration metrics_dict, images_dict = self.pipeline.get_eval_image_metrics_and_images(step=step) File "/workspace/nerfstudio/nerfstudio/nerfstudio/utils/profiler.py", line 93, in inner out = func(args, kwargs) File "/workspace/nerfstudio/nerfstudio/nerfstudio/pipelines/base_pipeline.py", line 329, in get_eval_image_metrics_and_images outputs = self.model.get_outputs_for_camera_ray_bundle(camera_ray_bundle) File "/opt/conda/lib/python3.8/site-packages/torch/autograd/grad_mode.py", line 27, in decorate_context return func(*args, kwargs) File "/workspace/nerfstudio/nerfstudio/nerfstudio/models/base_model.py", line 178, in get_outputs_for_camera_ray_bundle outputs = self.forward(ray_bundle=ray_bundle) File "/workspace/nerfstudio/nerfstudio/nerfstudio/models/base_model.py", line 140, in forward return self.get_outputs(ray_bundle) File "/workspace/nerfstudio/nerfstudio/nerfstudio/models/instant_ngp.py", line 180, in get_outputs ray_samples, ray_indices = self.sampler( File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl return forward_call(*input, *kwargs) File "/workspace/nerfstudio/nerfstudio/nerfstudio/model_components/ray_samplers.py", line 469, in forward ray_indices, starts, ends = self.occupancy_grid.sampling( File "/opt/conda/lib/python3.8/site-packages/torch/autograd/grad_mode.py", line 27, in decorate_context return func(args, kwargs) File "/opt/conda/lib/python3.8/site-packages/nerfacc/estimators/occ_grid.py", line 187, in sampling sigmas = sigma_fn(t_starts, t_ends, ray_indices) File "/workspace/nerfstudio/nerfstudio/nerfstudio/model_components/ray_samplers.py", line 414, in sigma_fn return density_fn(positions).squeeze(-1) File "/workspace/nerfstudio/nerfstudio/nerfstudio/fields/base_field.py", line 65, in densityfn density, = self.get_density(ray_samples) File "/workspace/nerfstudio/nerfstudio/nerfstudio/fields/nerfacto_field.py", line 243, in get_density h = self.mlp_base(positions_flat).view(*ray_samples.frustums.shape, -1) RuntimeError: cannot reshape tensor of 0 elements into shape [0, -1] because the unspecified dimension size -1 can be any value and is ambiguous