4eJIoBek1

Generfacto is dreamfusion but in nerfacto

https://docs.nerf.studio/nerfology/methods/generfacto.html Idk, you can read about generfacto here

Lol seems like you forgot to remove --medvram argument

to compile tinycudann you need to install CUDA and visualstudio c++ desktop module. After compiling you can use nerfstudio without cuda and visualstudio

You always need to compile tinycudann during installing

OMG use_fp16=False improves sampling speed dramatically! THX - 2 times faster sampling on gpu - 64 times faster sampling on cpu, with 20 steps you can make shapes for 10...

you should add to env variables HF_DATASETS_OFFLINE=1 TRANSFORMERS_OFFLINE=1 DIFFUSERS_OFFLINE=1 \ for conda just type conda env config vars set HF_DATASETS_OFFLINE=1 TRANSFORMERS_OFFLINE=1 DIFFUSERS_OFFLINE=1 \

Seems like no one is working :/

Oh, huge thanks