[bugfix] Add set_default_tensor_type to torch.DoubleTensor with precision=64

[tchaton]

What does this PR do?

c.f title.

Fixes #6927

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 🙃

[ethanwharris]

LGTM 😃

[codecov]

Codecov Report

Merging #7108 (be3a5b9) into master (3cc0b2c) will decrease coverage by 2%.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master   #7108    +/-   ##
=======================================
- Coverage      92%     90%    -2%     
=======================================
  Files         197     197            
  Lines       12634   12982   +348     
=======================================
+ Hits        11649   11697    +48     
- Misses        985    1285   +300     

[Borda]

shall you rather save what was it before and then set it back?

[tchaton]

I think it should be fine. Didn't find any utils to get the default either.

[ananthsub]

@tchaton this is breaking the plugin typechecks with mypy:

pytorch_lightning/plugins/precision/double.py:105: error: Incompatible types in assignment (expression has type "Callable[[], _GeneratorContextManager[Any]]", base class "Plugin" defined the type as "Callable[[], Generator[Any, Any, Any]]")  [assignment]
pytorch_lightning/plugins/precision/double.py:106: error: Incompatible types in assignment (expression has type "Callable[[], _GeneratorContextManager[Any]]", base class "Plugin" defined the type as "Callable[[], Generator[Any, Any, Any]]")  [assignment]
pytorch_lightning/plugins/precision/double.py:107: error: Incompatible types in assignment (expression has type "Callable[[], _GeneratorContextManager[Any]]", base class "Plugin" defined the type as "Callable[[], Generator[Any, Any, Any]]")  [assignment]
pytorch_lightning/plugins/precision/double.py:108: error: Incompatible types in assignment (expression has type "Callable[[], _GeneratorContextManager[Any]]", base class "Plugin" defined the type as "Callable[[], Generator[Any, Any, Any]]")  [assignment]

rather than overriding these step contexts, could we share the same implementation?