I've got a neural network model whose training would be sped up significantly if I could run it on my Nvidia GPU instead of my CPU. My model exists in my WSL2 environment -- WSL2 does not grant Tensorflow access to my computer's GPU.
As a workaround, perhaps the Python environment that is native to the Windows OS (as opposed to my WSL-Ubuntu install) has access to the GPU. However, runningpython.exe neuralnet.pyfrom WSL seems to exit immediately without output or error.
Is this workaround possible?