sethah opened a new pull request #9777: Mx 9588
   ## Description ##
   This PR adds a mixin class that F1 and other metrics like precision and 
recall can leverage in the future. It also provides a new option for the F1 
metric called `average` which defines how the metric will be aggregated across 
mini batches. 
   ## Checklist ##
   ### Essentials ###
   - [X] Passed code style checking (`make lint`)
   - [X] Changes are complete (i.e. I finished coding on this PR)
   - [X] All changes have test coverage:
   - Unit tests are added for small changes to verify correctness (e.g. adding 
a new operator)
   - Nightly tests are added for complicated/long-running ones (e.g. changing 
distributed kvstore)
   - Build tests will be added for build configuration changes (e.g. adding a 
new build option with NCCL)
   - [X] Code is well-documented: 
   - For user-facing API changes, API doc string has been updated. 
   - For new C++ functions in header files, their functionalities and arguments 
are documented. 
   - For new examples, is added to explain the what the example does, 
the source of the dataset, expected performance on test set and reference to 
the original paper if applicable
   - [X] To the my best knowledge, examples are either not affected by this 
change, or have been fixed to be compatible with this change
   ## Approach
   The "micro" vs "macro" update strategy is not specific to F1 score. The 
macro update just takes an average of averages, which can be done for any 
metric. It may be best to design an abstraction where any metric can have the 
micro/macro update option, but I couldn't see a good way to do that here that 
   * be easy to use for end users AND
   * maintain backward compatibility AND
   * maintain current semantics
   For now, the behavior for each type of update is hard coded into the 
`update` method of the `F1` class. We can discuss the approach.
   Please let me know if I have missed or overlooked anything :)

This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:

With regards,
Apache Git Services

Reply via email to