Installation of github repository using older CUDA toolkit

I have been trying to install Plenoxels (GitHub - sxyu/svox2: Plenoxels: Radiance Fields without Neural Networks) but installing it leads to errors and warnings related to CUDA. (CUDA is not detected when installing as recommended)

I now get past the point of errors I ran into first by installing a different version of CUDA toolkit (The Plenoxel environment doesn’t specify what version of CUDA toolkit to install but does specify what pytorch version to use - leading to incompatible versions and therefore a CPU-only install of pytorch)

However, I now run into another error: CUDA is not recognized when running the code after installation, but pytorch has GPU enabled. I think this is because I have to install CUDA toolkit using conda, but this does not result in it being recognized by

nvcc --version

(There is no folder in /usr/local/, so there is no environmental variable to be set) and running it just yields

Command ‘nvcc’ not found, but can be installed with:
sudo apt install nvidia-cuda-toolkit

Installing the correct version (one of the versions compatible with pytorch, 10.2 or 11.3) using either conda or pip still leads to an absent /usr/local/ folder or something that can be used to make nvcc -V work.

An example of (I think) the problem is cuda - Nvcc missing when installing cudatoolkit? - Stack Overflow.

The details of my machine and nvidia-smi for reference:

Ubuntu 22.04.3 LTS
| NVIDIA-SMI 535.129.03 Driver Version: 535.129.03 CUDA Version: 12.2 |
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
| 0 NVIDIA GeForce GTX 1070 Off | 00000000:01:00.0 On | N/A |
| N/A 60C P0 39W / 125W | 609MiB / 8192MiB | 1% Default |
| | | N/A |

Any help or insight is appreciated, please let me know if more information is required.