Config option(s) for ml devices

We should add an option to the worker config to choose the right device to be used for the ML tasks. I think at the moment only the pytorch-based models support this (i.e. pyannote for diariztion and torchaudio for alignment). https://github.com/transcribee/transcribee/pull/49#discussion_r1154541999

Start by locating the worker configuration and the PyTorch-based paths mentioned for pyannote diarization and torchaudio alignment. Trace how ML tasks choose their device and determine the supported configuration values. Done means the worker config can select the device used by those ML tasks, with the affected behavior verified.

Written by the indexing model from the issue text.