Running Pytorch on Intel dGPUs?

Is there anyway to run Pytorch on Intel dGPU, using oneAPI, or any other way?