Github user martinjaggi commented on the pull request:
https://github.com/apache/incubator-spark/pull/575#issuecomment-35217718
Hope you don't get me wrong, I was not at all proposing to fix a single
scheme, neither for serialization, or for the choice of sparse library. I was
just suggesting that the existing MLlib classification/regression code would be
a nice benchmark to see how the several candidate implementations perform in
reality (these only need vectors, no matrices). No matter what we will choose,
serialization time will also play an important role in the end.
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. To do so, please top-post your response.
If your project does not have this feature enabled and wishes so, or if the
feature is enabled but not working, please contact infrastructure at
infrastruct...@apache.org or file a JIRA ticket with INFRA.
