[bug] Set device on LightningModule after fitting

[tchaton]

What does this PR do?

This PR resolve a bug with device placement.
When running with gpu accelerator, the model was correctly put on cpu but the device was still set to cuda.

Before submitting

• Was this discussed/approved via a GitHub issue? (not for typos and docs)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

PR review

Anyone in the community is free to review the PR once the tests have passed.
Before you start reviewing make sure you have read Review guidelines. In short, see the following bullet-list:

• Is this pull request ready for review? (if not, please submit in draft mode)
• Check that all items from Before submitting are resolved
• Make sure the title is self-explanatory and the description concisely explains the PR
• Add labels and milestones (and optionally projects) to the PR so it can be classified

Did you have fun?

Make sure you had fun coding 🙃

[codecov]

Codecov Report

Merging #7188 (a5e226d) into master (c502e47) will decrease coverage by 5%.
The diff coverage is 50%.

@@           Coverage Diff           @@
##           master   #7188    +/-   ##
=======================================
- Coverage      92%     87%    -5%     
=======================================
  Files         197     197            
  Lines       12622   12667    +45     
=======================================
- Hits        11631   11000   -631     
- Misses        991    1667   +676     

[pep8speaks]

Hello @tchaton! Thanks for updating this PR.

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2021-04-23 13:45:29 UTC

[awaelchli]

LGTM.
Minor nit: perhaps the test could have been integrated in an existing test that just runs fit, in one of the tests/models/test_gpu tests perhaps.
but minor

[ethanwharris]

LGTM 😃

[ananthsub]

@SeanNaren for sharded training, do we need to move both self.model and self.lightning_module? Could this lead to errors like what we saw for eval/train with SDP earlier?