CUDA not working in conda installation

I installed Pytorch 1.3 with gpu support in my conda environment using the command given on the website. But, torch.cuda.is_available() is returning False. I already have a system wide install of CUDA 10.0(which I know is not for pytorch 1.3 but the sand-boxed 10.1 is already installed in my pytorch environment) and CuDNN 7.5.1 for TensorFlow 2.0.

On typing torch.version.cuda , I get 10.1.243 which is the sandboxed cuda toolkit installed with pytorch.

How do I resolve this version conflict? Please Help.

Thanks in advance.

Edit: Fixed the issue by updating the device driver.