Tensorflow-GPU in RAPIDS for Neural Networks to do Computer Vision

I figured that cuML does not have functions for Neural Networks, so I am wondering if it is possible to use tensorflow-gpu in RAPIDS. I installed tf-gpu in RAPIDS container but tf does not see the GPUs I checked it with


function and it returns FALSE. I am not sure what I am doing wrong.

Found out that we need to be using -dev version of the container.