Fixed implementation of pytorch cuda device

[LeandroCasiraghi]

We discovered the pytorch_device argument of get_sleep was not being passed on to sleepnet and hence never used.

We also fixed the sleepnet code to correctly implement the use of a cuda device in pytorch and added a printed message that indicates the device being used, either the default 'cpu' or the cuda device input in the get_sleep call.

[angerhang]

We discovered the pytorch_device argument of get_sleep was not being passed on to sleepnet and hence never used.

We also fixed the sleepnet code to correctly implement the use of a cuda device in pytorch and added a printed message that indicates the device being used, either the default 'cpu' or the cuda device input in the get_sleep call.

@LeandroCasiraghi thanks a lot for reporting this issue and making a PR.
The PR looks good to me.
Will merge it in once the tests finished :D

[angerhang]

We discovered the pytorch_device argument of get_sleep was not being passed on to sleepnet and hence never used.

We also fixed the sleepnet code to correctly implement the use of a cuda device in pytorch and added a printed message that indicates the device being used, either the default 'cpu' or the cuda device input in the get_sleep call.

There seems to be some format issues about the PR, which seems to be easy to change. Would you mind taking a look quickly :D @LeandroCasiraghi
https://github.com/OxWearables/asleep/actions/runs/5742758265/job/15566413256?pr=32

[LeandroCasiraghi]

fixed 5 format issues from tests

[angerhang]

Just a final formatting issue

[LeandroCasiraghi]

Hopefully final formatting issue