Skip to content

Issue of MonoSDF dataset format and scannet yaml config file #59

Open
@kevinchan04

Description

@kevinchan04

Hi, when I trying to train my custom dataset with monosdf dataloader, it has error as shown below:

Traceback (most recent call last):
File "code_single/tools/train.py", line 1699, in
main_function(bc.parse(print_config=False))
File "code_single/tools/train.py", line 1336, in main_function
image_patch_dataset = ImagePatchDataset(scene_dataloader_train, **params, **sampler_kwargs)
File "/srv/storage/kevin/Platform-NeRF_to_BIM/SourceCode/neuralsim/dataio/data_loader/image_loader.py", line 190, in init
self.sampler, self.scene_weights = get_frame_sampler(self.scene_loader, frame_sample_mode=self.frame_sample_mode, ddp=ddp, **sampler_kwargs)
File "/srv/storage/kevin/Platform-NeRF_to_BIM/SourceCode/neuralsim/dataio/data_loader/sampler.py", line 66, in get_frame_sampler
frame_weights = get_frame_weights_uniform(scene_loader, scene_weights, **kwargs)
File "/home/user/miniconda3/envs/nr3d/lib/python3.8/site-packages/torch/autograd/grad_mode.py", line 27, in decorate_context
return func(*args, **kwargs)
TypeError: get_frame_weights_uniform() got an unexpected keyword argument 'num_rays'

My dataset is built following monosdf format as https://github.com/autonomousvision/monosdf/blob/main/preprocess/scannet_to_monosdf.py . Not sure if it is suitable for neuralsim. Could you give any suggestions? Thank you.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions