Open Bardo-Konrad opened 3 months ago
Just as an FYI, tiny-cuda-nn is required for faster speedups. On Linux I have been able to avoid tcnn installation, and use nerfstudio (although almost 1.5x-2. slower).
Installation creates all kinds of errors because tinycudann does not work with CUDA Toolkit 12.1 and VS 12.10+, while VS 12.9. is not available anymore on windows. As this is not the repository of tinycudann, just replace it with something that's userfriendly on install.
for me tinnycudann is working with cuda 12.6
Installation creates all kinds of errors because tinycudann does not work with CUDA Toolkit 12.1 and VS 12.10+, while VS 12.9. is not available anymore on windows. As this is not the repository of tinycudann, just replace it with something that's userfriendly on install.