Github user hhbyyh commented on the issue:
https://github.com/apache/spark/pull/18924
Yes, I think local test is enough for both correctness and performance.
For consistency with old LDA, just some manual local test would be
sufficient. You may well just use the LDA example and switch the Spark jar
files between Spark 2.2 and your branch. And I think the case with empty
document worth special attention.
The same for the performance test. You may just post the result after your
local test. I'm OK as long as there's no noticeable regression.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
