remove optimizer_idx arg in manual optimization

[rohitgr7]

What does this PR do?

Removed optimizer_idx from training_step in case of manual optimization since it's redundant in such a case.

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 #6093 (947c2ef) into master (34b733b) will decrease coverage by 0%.
The diff coverage is 100%.

@@          Coverage Diff           @@
##           master   #6093   +/-   ##
======================================
- Coverage      94%     94%   -0%     
======================================
  Files         161     161           
  Lines       11474   11476    +2     
======================================
+ Hits        10731   10732    +1     
- Misses        743     744    +1     

[awaelchli]

@rohitgr7 can we warn that optimizer_idx is None and that it's not appropriate for manual optimization when the user still has it in the signature? For backward compatibility right?

[rohitgr7]

@rohitgr7 can we warn that optimizer_idx is None and that it's not appropriate for manual optimization when the user still has it in the signature? For backward compatibility right?

@awaelchli Can we deprecate the signature in this case with a warning and remove optimizer_idx? With the old one we can make it None. Wdyt?

[awaelchli]

Yes, None or whatever the current value is. I simply suggest to make both

def training_step(self, batch, batch_idx, optimizer_idx)
    ...

and

def training_step(self, batch, batch_idx)
    ...

valid signatures (for manual opt) so that switching to 1.3 will not break user code even if they ignored the argument so far. This will require signature inspection which is a bit ugly but I don't think we have another choice :)

[tchaton]

LGTM !

[awaelchli]

nice improvement