fix batch auto scaling when init_val causes OOM

[eladsegal]

What does this PR do?

When init_val batch size causes OOM, low variable is never set but is still required in the exception handling.
This pull request gives low a default value of 1 which will be overwritten if init_val doesn't cause OOM.

Does your PR introduce any breaking changes? If yes, please list them.

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 list all the breaking changes introduced by this pull request?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

PR review

Anyone in the community is welcome to review the PR.
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 #8954 (54ae6dc) into master (de22e40) will decrease coverage by 4%.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master   #8954    +/-   ##
=======================================
- Coverage      93%     89%    -4%     
=======================================
  Files         176     176            
  Lines       14270   14271     +1     
=======================================
- Hits        13260   12681   -579     
- Misses       1010    1590   +580     

[awaelchli]

thanks for the fix.
this is caused when we get OOM at the first trainer run, correct?

[eladsegal]

thanks for the fix.
this is caused when we get OOM at the first trainer run, correct?

Correct!

[awaelchli]

@eladsegal could you go through the checklist on the PR description and add a changelog entry too?
I believe we can omit the test due to its complexity and dependence on the particular system and load.

[eladsegal]

Sure, done @awaelchli

[tchaton]

LGMT !