search

bennyguo / instant-nsr-pl

Neural Surface reconstruction based on Instant-NGP. Efficient and customizable boilerplate for your research projects. Train NeuS in 10min!
MIT License
856 stars 84 forks source link

issue when training on colmap data #82

Closed huang997733 closed 1 year ago

huang997733 commented 1 year ago

Hi there,

I have encountered this issue when training on colmap data (neus-colmap and nerf-colmap)

RuntimeError: indices should be either on cpu or on the same device as the indexed tensor (cpu)

This error message shows up when the training process hits 5000 iterations I have searched online but nothing really helpful Any ideas how to solve this issue?

Thanks so much!

Yellowshuohahaha commented 1 year ago

Hi there,

I have encountered this issue when training on colmap data (neus-colmap and nerf-colmap)

RuntimeError: indices should be either on cpu or on the same device as the indexed tensor (cpu)

This error message shows up when the training process hits 5000 iterations I have searched online but nothing really helpful Any ideas how to solve this issue?

Thanks so much!

I have the same problem. Have you solved it?

huang997733 commented 1 year ago

Hi there, I have encountered this issue when training on colmap data (neus-colmap and nerf-colmap) RuntimeError: indices should be either on cpu or on the same device as the indexed tensor (cpu) This error message shows up when the training process hits 5000 iterations I have searched online but nothing really helpful Any ideas how to solve this issue? Thanks so much!

I have the same problem. Have you solved it?

This is because the dataset is not loaded on gpu, simply change load_data_on_gpu in configs/neus-colmap.yaml to true would fix the problem. It took me several hours to find out lol.

KeKer7 commented 4 months ago

Can you train Neus on the Colmap dataset and achieve good results? I can't render it on the mip nerf dataset and tTAT dataset, it always doesn't converge