MrForExample/ComfyUI-3D-Pack

Notes about compiling the dependencies

doctorpangloss opened this issue ยท 3 comments

These instructions can be adapted to compile your requirements and avoid using miniconda on Windows.

# using git bash for the sake of simplicity
# enable developer mode
# google this: allow os.symlink on windows by adding your username to the local security policy entry for it.
# you will have to restart your computer
# install chocolatey using powershell, then install the prereqs for compilation on Windows
choco install -y visualstudio2022buildtools
choco install -y visualstudio2022-workload-vctools --package-parameters "--add Microsoft.VisualStudio.Component.VC.Llvm.ClangToolset --add Microsoft.VisualStudio.Component.VC.Llvm.Clang"
choco install -y cuda

# now clone threestudio
git clone git@github.com:threestudio-project/threestudio.git
cd threestudio/
python -m venv venv
source venv/scripts/activate
pip install torch torchvision --index-url https://download.pytorch.org/whl/cu121
# include missing dependencies from their repo
pip install ninja wheel mediapipe
pip install -r requirements.txt --no-build-isolation
# you will watch nerfacc and tiny-cuda compile for a while
# download stable_zero123.ckpt from https://huggingface.co/stabilityai/stable-zero123 and put it into ./load/zero123
# run. modify your configs/stable-zero123.yaml data.random_camera.batch_size, trainer.precision and trainer.gradient_accumulation_steps for much better performance
python launch.py --config configs/stable-zero123.yaml --train --gpu 0 data.image_path=./load/images/hamburger_rgba.png

Thanks friend.
So from what I can tell, this is a windows only solution, and basically using one package manager (Chocolatey) to replace another (Miniconda).
But still, thanks for a different solution, I'll add it into install guide soon since some user may find this solution helpful :)
Cheers ๐Ÿ‘

Thanks friend.

So from what I can tell, this is a windows only solution, and basically using one package manager (Chocolatey) to replace another (Miniconda).

But still, thanks for a different solution, I'll add it into install guide soon since some user may find this solution helpful :)

Cheers ๐Ÿ‘

lol okay good point. But installing visual studio via its setup is really complicated. This would all work with the embedded Python.

chocolatey is comparable to apt not miniconda. Overall blame Microsoft.

where I build this venv