NVIDIAGameWorks / kaolin-wisp

NVIDIA Kaolin Wisp is a PyTorch library powered by NVIDIA Kaolin Core to work with neural fields (including NeRFs, NGLOD, instant-ngp and VQAD).
Other
1.45k stars 132 forks source link

Installation with CUDA 11.2 #120

Closed LouisSerrano closed 1 year ago

LouisSerrano commented 1 year ago

Is it possible to install wisp for CUDA 11.2 ?

orperel commented 1 year ago

Hi @LouisSerrano , CUDA 11.2 is supported, you can follow the manual installation for kaolin (mostly git clone and python setup.py develop: https://kaolin.readthedocs.io/en/latest/notes/installation.html#installation-from-source

If you're looking for a binary installation, have a look at the kaolin wheels matrix here: https://kaolin.readthedocs.io/en/latest/notes/installation.html#quick-start-linux-windows kaolin released binary wheels for CUDA 11.1 or CUDA 11.3, so alternatively you can opt for either versions instead.

The wisp installation should remain the same for all kaolin versions.

LouisSerrano commented 1 year ago

Ok thanks, I'll try to do this :)

LouisSerrano commented 1 year ago

Ok I had to install CUDA 11.7 to install it properly, otherwise I had a warning with incompatible torch and cuda version during the installation of wisp. Thanks, it seems to work now !