search

ToniRV / NeRF-SLAM

NeRF-SLAM: Real-Time Dense Monocular SLAM with Neural Radiance Fields. https://arxiv.org/abs/2210.13641 + Sigma-Fusion: Probabilistic Volumetric Fusion for Dense Monocular SLAM https://arxiv.org/abs/2210.01276
BSD 2-Clause "Simplified" License
1.18k stars 144 forks source link

vio-slam #14

Closed liangyongshi closed 1 year ago

liangyongshi commented 1 year ago

I0103 15:19:20.003431 146533 pipeline_module.py:85] Module: VioSLAM - Spinning. Segmentation fault (core dumped)

ToniRV commented 1 year ago

Ah! Can you provide a stack trace? perhaps with gdb?

liangyongshi commented 1 year ago

ubuntu 20.04, torch 12.1, CUDA 11.6, 1 RTX3090, I have only one GPU, so modify the code as the picture modified run the code python ./examples/slam_demo.py --dataset_dir=./Datasets/Replica/office0 --dataset_name=nerf --buffer=100 --weights=./droid.pth --img_stride=2 --fusion='nerf' --slam --gui --parallel_run

results result

liangyongshi commented 1 year ago

when I close the parallel_run, the result is Screenshot from 2023-01-05 17-33-46

Looking forward to your reply

ZhenhuiL1n commented 1 year ago

I am also facing this problem, did anyone solve it yet?

ZhenhuiL1n commented 1 year ago

bug

ZhenhuiL1n commented 1 year ago

bug1

liangyongshi commented 1 year ago

I am also facing this problem, did anyone solve it yet?

something wrong with torch_scatter and droid_backends, adopt conda to insatll torch_scatter, and I don't know how to solve the other problem of droid_backends

ZhenhuiL1n commented 1 year ago

I am also facing this problem, did anyone solve it yet?

something wrong with torch_scatter and droid_backends, adopt conda to insatll torch_scatter, and I don't know how to solve the other problem of droid_backends

I think it should be a specific version, cause I used the conda to install troch_scatter and other bug pops up

phamquandung commented 1 year ago

For me, I use two GPU and the problem is solved