rcurtin_irc changed the topic of #mlpack to: mlpack: a scalable machine learning library (https://www.mlpack.org/) -- channel logs: https://libera.irclog.whitequark.org/mlpack -- NOTE: messages sent here might not be seen by bridged users on matrix, gitter, or slack
<NabanitaDash[m]> https://github.com/mlpack/mlpack/wiki/DesignGuidelines#inheritance I vaguely understand why we shouldn't use inheritance with template functions. As per my understanding, the only reason we don't use inheritance in design features is due to memory overload. So, can we use it where memory overload is negligible? Can somebody please explain it a little bit more?
<jonpsy[m]> Big day tommorow! :-)
<shashankshet[m]> Is tomorrow's meeting for GSoC mentors only?
<shashankshet[m]> Or is it a meeting for both candidates and mentors?
<jonpsy[m]> These meetings have always been open for all.
<shashankshet[m]> Okay, and any idea what time?
<jonpsy[m]> have a look at mlpack/community.
<shashankshet[m]> Oh okay,
<shrit[m]> 17 UTC
<NabanitaDash[m]> s/```/``/, s/```/``/
baracurda_ClassN has joined #mlpack
baracurda_ClassN has quit [Remote host closed the connection]
baracurda has joined #mlpack
<zoq[m]1> <NabanitaDash[m]> "https://github.com/mlpack/mlpack..."; <- Usually we avoid virtual functions because the vtable lookup is slow (not 100% true anymore), so if you do something very often in a loop it has a performance penalty. That said we realized that for some methods like the Neural Network code it makes the design a lot cleaner and the vtable slowdown is neglectable compared to the main computation, so we are going to use inheritance for at
<zoq[m]1> least that part.
<zoq[m]1> zoq[m]1: It would be possible to use it for other parts as well, but we should discuss that for each case and check if we can use another solution as well.
<NabanitaDash[m]> That sounds about right, thanks!
baracurda has quit [Ping timeout: 276 seconds]
kristjansson has quit [Ping timeout: 256 seconds]
kristjansson has joined #mlpack