Failed to run model on GPU

I am using Pytorch on a remote machine provided by university.

I am trying to train my model using GPU, but unable to do so because of some issues with cuda.

Even though cuda is available, I am unable to use it.

Any help will be appreciated.

Attached screenshot for reference

An this is the GPU specification

It looks like python is utilizing almost all of your gpu memory. Is this a shared machine? Maybe someone else is using it.

Yes, it worked on different machine. Thanks though