ComfyUI 3D Pack.md

How to setup a Conda/Pytorch enviroment for using ComfyUI's 3D-Pack

Spoiler allert the solution was to target CUDA 11.8 (PyTorch 2.2.0+)

1. Make sure you have git, git-lfs, mesa and CUDA-11-8 installed at a OS level

sudo apt install git git-lfs cuda-11-8 mesa-utils mesa-common-dev libegl1-mesa-dev libgles2-mesa-dev 

2. Install conda

3. Create a new conda enviroment with Python 3.10, CUDA 11.8 and PyToch 2.2.0

conda create -n comfy3D python=3.10
conda activate comfy3D
pip install -U pip
conda install cuda -c nvidia/label/cuda-11.8.0
conda install pytorch=2.2.0 torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
conda install pytorch3d -c pytorch3d
pip install kiui 
pip install git+https://github.com/NVlabs/nvdiffrast/
pip install torch-scatter -f https://data.pyg.org/whl/torch-2.2.0+cu118.html 

4. Install ComfyUI (if you haven't)

git clone [email protected]:comfyanonymous/ComfyUI.git
cd ComfyUI
pip install -r requirements.txt

5. Install 3D Pack

cd custom_nodes
git clone -v --recursive --progress -- https://github.com/MrForExample/ComfyUI-3D-Pack
cd ComfyUI-3D-Pack
pip install -r requirements.txt