CUDA Issues running NeRF example

[javieryu]

Running into some issues running the nerf example.

My setup is:
OS: Ubuntu 20.04
GPU: RTX 3070 Ti
CUDA: 11.6 (driver 510)

Pretty sure CUDA is install correctly and I've added these lines to my .zshrc:
export PATH="/usr/local/cuda-11.6/bin:$PATH"
export LD_LIBRARY_PATH="/usr/local/cuda-11.6/lib64:$LD_LIBRARY_PATH"

[ashawkey]

Hi, this seems not to be a problem of this repo. Could you install other pytorch extensions, like the tiny-cuda-nn (pip install git+https://github.com/NVlabs/tiny-cuda-nn/#subdirectory=bindings/torch)?
A related issue here maybe helpful.

[javieryu]

Solved, I was trying to run the code with CUDA related flags that were not compatible with my GPU. Removing the CUDA_VISIBLE_DEVICES=1 from the run_nerf.sh script solved this issue.