How to make a Virtual Linux Machine see GPU on a Windows Host machine on WSL2 or VirtualBox

Im trying to run a Computational Fluid Dynamics software that requires GPU connection on Linux! I do have a Windows Host system and tried to intall WSL2 in order to have Linux running. Afterward I tried the Virtual machine (Virtualbox) and none of the cases I was able to see the GPU while on Linux terminal! Im with a GTX 1660 TI on my laptop. It is a Dell laptop with Windowns 10 Home Insider Build 20201.
I tryied different foruns but looks that nobody knows how to do it!
Can someone from NVIDIA help me on this subject!
I`m sure that this topic will help a lot of engineers interest in accelerate Matlab and Python as well, so please NVIDIA advise!
Thanks

I think it is more likely you will get a response to this question in the sub-forum dedicated to CUDA on WSL2:

https://forums.developer.nvidia.com/c/accelerated-computing/cuda/cuda-on-windows-subsystem-for-linux/303