[#1804] Re-implement pytorch profiler pr

[florianscheidl]

Description

After #1804 was reverted, we slightly rewrote the profiler PR, which should avoid the issue with the profiling config.

✅ Important: wait for #2310 (review) to be merged first.

Issue Number

Fixes #1804

Is this PR a draft? Mark it as draft.

Checklist before asking for review

• I have performed a self-review of my code
• My changes comply with basic sanity checks:
  • I have fixed formatting issues with ./scripts/actions.sh lint
  • I have run unit tests with ./scripts/actions.sh unit-test
  • I have documented my code and I have updated the docstrings.
  • I have added unit tests, if relevant
• I have tried my changes with data and code:
  • I have run the integration tests with ./scripts/actions.sh integration-test
  • (bigger changes) I have run a full training and I have written in the comment the run_id(s): launch-slurm.py --time 60
  • (bigger changes and experiments) I have shared a hegdedoc in the github issue with all the configurations and runs for this experiments
• I have informed and aligned with people impacted by my change:
  • for config changes: the MatterMost channels and/or a design doc
  • for changes of dependencies: the MatterMost software development channel

[grassesi]

I have not run it yet, but I am pretty sure there are still some configuration issues. Please see the comments for more details. I am also happy to discuss the implementation details I suggested.