Hey,
I'm Harsh Patel, and I am looking to contribute to MXNet. I wanted to get
some feedback to improvements that could be made with the current structure
that we have for automatically labelling issues and pull requests. I have
linked my proposed design structure on the bottom of this wiki page (
https://cwiki.apache.org/confluence/display/MXNET/Machine+Learning+Based+GitHub+Bot)
- it should be under 7. Overall, users will benefit from this design since
it will allow adding, updating, and deleting of labels freely. Label
creation will be faster since this model focuses on labelling an issue as
soon as it is made. Another key benefit is that we minimize the number of
total GitHub API calls that need to be made. Feedback would be much
appreciated - I would like to hear what the developers have to say! Thanks.
