ChanServ changed the topic of #mlpack to: "mlpack: a fast, flexible machine learning library :: We don't always respond instantly, but we will respond; please be patient :: Logs at http://www.mlpack.org/irc/
< RishabhGargGitte> Hey, @rcurtin, I received your comments and they were very helpful. I have responded to them and I need a few clarifications from you in some of them. Please do take a look when you get time. Thanks :)
< jonpsy[m]> What's up with ensmallen static build check though?
ImQ009 has joined #mlpack
ImQ009 has quit [Read error: Connection reset by peer]
ImQ009 has joined #mlpack
< RishabhGargGitte> @rcurtin, I have compiled all of my thoughts about the refactoring for regression trees into [this](https://docs.google.com/document/d/146WZ155o0OxMzEWIIbqQxJ844XE_lfqTUWh4fRJrkR8/edit?usp=sharing) doc. I have done quite some thinking about each of them. It would be great and I can start working on it immediately if you could take a look at it and see if my ideas are flowing in the right direction. Thanks!
< RishabhGargGitte> The refactoring would take quite some time and effort to be done, so I just want to be sure that my approach is good enough :)
ImQ009 has quit [Quit: Leaving]
ImQ009 has joined #mlpack
ImQ009 has quit [Read error: Connection reset by peer]
ImQ009 has joined #mlpack
< zoq> jonpsy[m]: I'm updating the build pipeline so it's easier to maintain. Once #284 is merged at least ensmallen will be back to normal.
< jonpsy[m]> You're adding sanity checks if im right..? And changing some files to adhere to those checks.
< jonpsy[m]> My job in NSGA2 is to fix the failing sanity check(to be merged) if it comes. Right?
< zoq> You don't have to, I can fix the warning as well.
< zoq> Pretty minor issues.
< jonpsy[m]> Ah init issues, so if you handle. I'll just rebase it, that'll do the trick?
< zoq> Might be easier to submit a patch to the open PR, I guess that will avoid some merge conflicts later on.
< jonpsy[m]> Yeah!
< zoq> I can do that.
< jonpsy[m]> If you insist
< zoq> Nah, if you like to fix the warning, please feel free.
< jonpsy[m]> :D
< jonpsy[m]> Gimme a min I'll do it rn
< zoq> Nice, thanks!
< zoq> Now I just have to bribe someone to approve the other PR :)
< rcurtin[m]1> no bribe necessary, I just need these meetings to end 😃
< rcurtin[m]1> my company is having a "virtual con", so all meetings all week long
< rcurtin[m]1> it's not bad, it's just... a lot
< zoq> so you are back to meeting across the entire week, if I remember right you somewhat managed to have all meetings on a single day.
< zoq> I guess lucky you it's just this week.
< OleksandrNikolsk> how do you manage working on an open source project + in a company?
< OleksandrNikolsk> It's intersting for me, in terms of "work-life-balance"
< jonpsy[m]> <zoq "Now I just have to bribe someone"> Where can I sign up? Asking for a friend of course :)
< zoq> jonpsy[m]: hehe, I'll approve the NSGA PR once we get a green build, which I think is what we will get once the other one is merged.
< jonpsy[m]> Can we wait for recursive Arbitrary check. It's really really close!
< zoq> Same for #283 :)
< jonpsy[m]> Yeah, let's do static fix then 283 and then NSGA2
< jonpsy[m]> Man, tonight's gonna be a blast
< zoq> Sounds good
< rcurtin[m]1> Oleksandr Nikolskyy: good question... when I was in graduate school, my thesis work was built on top of mlpack so it was easy
< zoq> It will proably be another 24 hours.
< rcurtin[m]1> now it is a little bit harder, because my work at my company does use mlpack, but I can't work on mlpack full-time there
< rcurtin[m]1> so, I try to balance the mlpack work with free time... typically I try to put ~1-1.5 hours per day into mlpack---sometimes more if I'm having fun---so that I can also enjoy other things and hobbies
< rcurtin[m]1> but, it's tough, because limiting my time into mlpack means I can't be on top of all PRs or issues, and sometimes there are emails that I simply am not able to respond to, which is not a good feeling
< rcurtin[m]1> this talk is very intro level, so I was able to devote some quick attention to #284 to approve it :)
< OleksandrNikolsk> I see
< OleksandrNikolsk> Just wondering what the future (after graduating) can look like lol
< OleksandrNikolsk> I'm studying+working part time as web dev (which is becoming boring). Have to take free time from job to check out open source, which also means - less money
< rcurtin[m]1> yeah; one option, it is possible to make a living through open source; for instance, one could, e.g., consult with companies that need data science tasks completed, and then use mlpack for that work / also develop mlpack for those needs
< rcurtin[m]1> another option is to find a company that is interested in the functionality of the open source project and hires you to build upon it
< rcurtin[m]1> when I worked at Symantec, that was their deal with me---half my time to work on mlpack, half my time deploying it for their needs
< OleksandrNikolsk> sounds cool
< OleksandrNikolsk> also interesting to hear about usage of mlpack by companies
< rcurtin[m]1> and, at the very least, open source leads to some amount of notoriety; saying that I maintain mlpack and develop for it seems to be something that is very attractive for people who are looking for developers to hire
< rcurtin[m]1> of course, those are only my experiences; maybe I am lucky? I have no way to know; I only have my own experiences to share 😃
< OleksandrNikolsk> thanks for sharing !
< OleksandrNikolsk> Now it's time for me to go back working on my proposal lol
< rcurtin[m]1> 👍️
< jonpsy[m]> i dont know how i did it, but i somehow got half the commits before master and half the commits after master and force pushed it. Yayy
< OleksandrNikolsk> surviving in the git jungle
< jonpsy[m]> * Bee Gees : Staying Alive plays in background *
< OleksandrNikolsk> If I copy-paste algorithms from papers to my proposal, should I provide references?
< OleksandrNikolsk> I mean pseudocode
< AnmolpreetSinghG> Hi @zoq, thanks for your comments. I have wrote my plans for using existing functionality and need of additional functions for image processing specially Gaussian filter 2d convolution, in reply to your comment in draft proposal. Kindly share your thoughts on that, it will help in further planning.
ImQ009 has quit [Quit: Leaving]
pavol_m[m] has joined #mlpack
< rcurtin[m]1> shrit: I got the arm64 tests to all run successfully with a fix to the cover tree, but some of them still fail, so clearly I have more to debug :)