For Windows Users

[jtydhr88]

if you try to run this on Windows, you might need to install the following dependencies manually:

1. conda install -c conda-forge cudatoolkit-dev -y
2. pip install Ninja
3. https://visualstudio.microsoft.com/visual-cpp-build-tools/
   and install c++ build tools with
   

after installed these additional dependencies, I could run it on my local with 3090

Great Job!

[Mikerhinos]

What is the VRAM consumption of the tool ?

[Lektro9]

What is the VRAM consumption of the tool ?

Using Win10, Docker and 3090 - it's using 23.3 GB VRAM (could be a little less since I am running some other programs)

[abuneo]

what is conda? i found condamini should i install that?

also if i only have a 3060 will i be able to run it and get good res?