Grid K1 and CUDA for deep learning

Hi, I have a Nvidia Grid K1 card using with ESXi 6.5 vGPU on a Windows 7 Server, using K180Q mode. I would like to use it for deep learning with tensorflow and python.

I am using the latest driver pack: NVIDIA-vGPU-kepler-vSphere-6.5-367.134-370.41, and using 370.41_grid_win7_win8_server2008R2_serveur2012R2_64bit_international.exe for installing the driver.

I also followed the instructions on GPU support  |  TensorFlow
and downloaded the CUDA version 10.1 as I’m using Tensorflow 2.1.0.
I also download CuDNN version 7.6.4 for CUDA.
Everything looks correctly installed and I added all the correct paths in the PATH (Environment variable).

But as I try the command tensorflow.test.is_gpu_available() in the command prompt saying “Failed call to cuInit: CUDA_ERROR_NO_DEVICE: no CUDA-capable device is detected”

Any ideas on how to get around that ?
Thank you