mlpack IRC logs, 2020-03-19

Logs for the day 2020-03-19 (starts at 0:00 UTC) are shown below.

March 2020
--- Log opened Thu Mar 19 00:00:34 2020
04:54 -!- hj [74497b81@] has joined #mlpack
04:55 < Param-29Gitter[m> @naruarjun have a look at implementation of following algorithms
04:55 < Param-29Gitter[m> 1) K-means
04:55 < Param-29Gitter[m> 2) Random forest.
04:58 -!- hj [74497b81@] has quit [Remote host closed the connection]
05:52 -!- atreya [~atreya@] has joined #mlpack
05:54 -!- atreya [~atreya@] has quit [Client Quit]
05:56 < TanayMehtaGitter> Hey everyone! I was working on my proposal, but I wondered if it's OK to send a draft (for reviewing) to the respective mentor personally?
07:14 < Param-29Gitter[m> Hey @rcurtin , @zoq will we be having video meetup today (1800 UTC on Thursday )? It has been 2 weeks :).
07:15 < Saksham[m]> yeah we received a mail regarding it
07:35 < jeffin143[m]> Yes I am in
07:40 < chopper_inbound4> For some reason I am always absent in video chats :( I am currently on my way to home with bad internet connection.
07:45 < jenkins-mlpack2> Project docker mlpack nightly build build #646: FAILURE in 3 hr 31 min:
07:46 < Param-29Gitter[m> Can someone send link of the message. I didn't receive it.
07:47 < Saksham[m]> Tomorrow’s the third Thursday of the month, so let’s have another video
07:47 < Saksham[m]> call at 1800 UTC.
07:47 < Saksham[m]> <>
07:47 < jeffin143[m]> :)
07:48 < Param-29Gitter[m> Thanks :)
07:49 < jeffin143[m]> Any clue where is Toshal Agrawal (Gitter) ?
07:50 < himanshu_pathak[> Saksham: I have questions about the project :) will ask there
07:51 -!- Heisenbug [7c42adce@] has joined #mlpack
07:52 < jeffin143[m]> Also sreenik :)
07:52 < Heisenbug> Hello, regarding GSoC I want to have a discussion regarding the Visualization Tool Project.
07:53 < Heisenbug> I also mailed on mailing list but there was no response and as date is coming closer I need some insight from mentors to prepare for the project proposal.
07:53 < Heisenbug> I have already started to working on the tool, and made some developed some basic functionalities.
07:54 < jeffin143[m]> Hi Heisenbug :)
07:54 < jeffin143[m]> I was working on the same proposal
07:54 < Heisenbug> I plotted accuracy curves for models in model zoo, I also visualized K-Means, which shows real time clustering...
07:54 < jeffin143[m]> What have used as bare tool ??
07:54 < Heisenbug> Hiii jeffin143[m]
07:55 < jeffin143[m]> Have you build it or using something ?
07:55 < Heisenbug> OpenCV and Matplotlibcpp
07:55 < jeffin143[m]> Is it a dashboard , or saving it as file ??
07:56 < Heisenbug> We can save it as a file...
07:56 < Heisenbug> Shall I open a PR for kmeans?
07:56 < Heisenbug> so, that you can see it by yourself?
07:56 < jeffin143[m]> We were looking for a dashboard thing , so that user can see real time action
07:56 < jeffin143[m]> Something similar to websocket
07:57 < jeffin143[m]> Like tensorboard , where as you are training the graph gets updates
07:57 < Heisenbug> It is real-time, you can see the clustering...
07:57 < Heisenbug> Yeh, it will open a window and will train realtime...
07:58 < jeffin143[m]> Oh really
07:58 < jeffin143[m]> Ok so basically you are plotting it again after every small intervals , right ?
07:58 < Saksham[m]> jeffin143 do you any idea about the CMAKE project ?
07:59 < Heisenbug> yea, the cluster points will be plotted in different colors for different clusters...Each point will be plotted as the algo trains...
07:59 < jeffin143[m]> @Saksham if I am not wrong is the mentor of the project
08:00 < jeffin143[m]> @Heisenbug : did you submit your draft proposal
08:00 < jeffin143[m]> ?
08:01 < jeffin143[m]> Saksham : idea in the sense , are you going to mentor that ?
08:01 < Manav-KumarGitte> I wanted to have discussion on application of ANN implemented in mlpack project. Can somebody help me with it..
08:01 < Saksham[m]> I think you are confusing me for the wrong saksham 😛
08:02 < jeffin143[m]> Ohh
08:02 < jeffin143[m]> Exactly
08:02 < jeffin143[m]> Same name
08:02 < jeffin143[m]> My bad
08:02 < Heisenbug> No I didn't jeffin143[m]...
08:02 < Saksham[m]> which projects are you mentoring ?
08:03 < jeffin143[m]> Ok just go ahead , record a video of the clustering thing and please pitch your idea today , just spare some time and put up the proposal so that mentors can see
08:03 < Heisenbug> Shall I submit one?
08:03 < jeffin143[m]> Yes heisenbug , go ahead and submit it
08:03 < jeffin143[m]> Asap
08:04 < jeffin143[m]> Saksham: I will be trying the gsoc as a student :) most probably
08:04 < jeffin143[m]> If that doesn't work out I will still be helping here and there and learning things out :)
08:05 < Heisenbug> Which project are you working on?
08:06 < jeffin143[m]> I was working on the same project as you , the dashboard one
08:06 < Heisenbug> was working? Are you not now?
08:07 < jeffin143[m]> So I thought of using boost::asio to transfer the logs and all the information to web frontend so that it can all been seen at once
08:07 < jeffin143[m]> But rcurtin suggested me to find an alternative
08:08 < jeffin143[m]> Since building a full dashboard in 3 months in not realistic
08:08 < jeffin143[m]> So he told me to integrate it with some existing dashboard
08:08 < jeffin143[m]> I did research it a bit
08:08 < jeffin143[m]> I am not sure how would we really do it
08:09 < Heisenbug> But the idea itself says that integrating will increase the dependency and will also not have all the functionalities we need.
08:09 < jeffin143[m]> Heisenbug: if you have to plot 3 graphs simultaneously , did it open 3 windows or a window with 3 pane
08:09 < Heisenbug> I haven't tried it out yet, I am just plotting one at a time...
08:09 < Saksham[m]> not working on this project, but would wrapping something like tensorboard not be a good idea?
08:10 < Heisenbug> That would be good Saksham[m] but we can't build that directly, it would take a lot of time for that.
08:10 < jeffin143[m]> <Heisenbug "But the idea itself says that in"> Heisenbug: tensorboard has all the functions we need ,
08:11 * jeffin143[m] sent a long message: < >
08:11 < jeffin143[m]> And I have very less time to do that 😂
08:11 < jeffin143[m]> May be before 31st march
08:11 < Heisenbug> I dont think personally thats a good idea...
08:12 < jeffin143[m]> <Heisenbug "That would be good Saksham but w"> Heisenbug: yes it will the main reason behind , our Ann infra is full c++ , I mean we can easily do it for machine learning algorithms
08:12 < Heisenbug> what i want is, build a class Visualise which will have all the code for it now, complete it this summer...
08:12 < jeffin143[m]> And that is the where majority of work is being done and hence without Ann support , it wouldn't be match of helpp
08:12 < jeffin143[m]> Ann in the sense lstm gan and everything
08:12 < Heisenbug> and then build a front-end for it later...
08:13 < Heisenbug> we can use the pre existing code...
08:13 < Heisenbug> we then just need to build a front-end which will help users navigate through the graphs or they can directly access it from code.
08:14 < Heisenbug> I have also plotted curves for validation accuracy vs training accuracy...
08:14 < jeffin143[m]> Heisenbug , if you using plotting the graph in window
08:14 < jeffin143[m]> How are you planning to redirect it to frontend m
08:14 < jeffin143[m]> ?
08:14 < Heisenbug> it will update after each epoch...
08:14 < Heisenbug> That what I want to discuss with mentors...
08:16 < PrinceGuptaGitte> Hi, sorry to interrupt in between a conversation. I had a little doubt about GSOC proposal. Should I mention only the currently open/merged PRs or should I also mention the ones which got closed in Contributions section.
08:16 < jeffin143[m]> Prince Gupta (Gitter): you can mention all the pr :)
08:16 < jeffin143[m]> Closed opened merged
08:16 < jeffin143[m]> All repo :)
08:17 < PrinceGuptaGitte> Ok I will. Thanks for telling me. :)
08:17 < jeffin143[m]> Heisenbug: anyways just go ahead and submit the proposals, They would get back to you
08:18 < jeffin143[m]> Isn't opencv and matplotlibcpp a dependency then ??
08:19 < Saksham[m]> does gsoc allow us to edit our proposals ?
08:19 < Heisenbug> Yea, I know...but we can atleast build the whole tool as we want...
08:20 < Heisenbug> It was also mentioned in the idea that we can use opencv so lets see...
08:20 < Heisenbug> Yea, we can edit...
08:20 < Heisenbug> we can edit only draft proposals..
08:20 < jeffin143[m]> Saksham ; draft proposal are online docs so you can edit them
08:21 < jeffin143[m]> But final is pdf
08:21 < Heisenbug> You had discussion with the mentor for this project jeffin143[m]?
08:26 < jeffin143[m]> No I just had a small suggestions from rcurtin
08:26 < jeffin143[m]> I haven't be able to get in touch with ToshalAgrawal[m] about it
08:28 < Heisenbug> Ryan Birmingham is the mentor for this project right?
08:28 < Heisenbug> I also mailed on mailing list regarding this but no reply till now...
08:31 -!- Heisenbug [7c42adce@] has quit [Remote host closed the connection]
08:33 -!- heisenbug [7c42adce@] has joined #mlpack
08:34 < jeffin143[m]> Heisenbug: which project
08:34 < jeffin143[m]> ?
08:34 < heisenbug> Visualisation Tool
08:34 < heisenbug>
08:36 < jeffin143[m]> Oh ok I just had a look on it
08:37 < jeffin143[m]> I wasn't aware he is the mentor :)
08:37 < heisenbug> No worries, that check that out...
08:37 < heisenbug> maybe mentors will join in here...
08:37 < jeffin143[m]> :)
08:40 -!- heisenbug is now known as heisenbuug
08:42 < Manav-KumarGitte> @rcurtin the cmake flag you suggested me didn't worked for building on Mac OS. I am really struggling with it. Do you have any other alternative?
08:44 < vigsterkr[m]> Manav-Kumar (Gitter): check their CI definitions in the repository for macos
08:44 < vigsterkr[m]> you can learn a lot of things from there
08:44 < Saksham[m]> yeah
08:45 < vigsterkr[m]>
08:45 < Saksham[m]> i followed it and seemed to work perfectlly
08:46 -!- favre49 [~favre49@] has joined #mlpack
08:47 < Manav-KumarGitte> Ok, i follow it.
09:11 < Manav-KumarGitte> It's still not working. Getting the same error as before. I was supposed to just use the system default compiler and i have done that but not working
09:57 -!- ImQ009 [~ImQ009@unaffiliated/imq009] has joined #mlpack
09:59 -!- heisenbuug [7c42adce@] has quit [Remote host closed the connection]
11:26 * hemal[m] sent a long message: < >
11:28 < Manav-KumarGitte> Are you too on Mac OS?
11:28 < hemal[m]> no, linux ubuntu
11:29 < Manav-KumarGitte> ohh, i am facing the same problem from last few days
11:29 < hemal[m]> any workaround ?
11:29 < Manav-KumarGitte> On Mac
11:31 < Manav-KumarGitte> They just asked to follow the ci definitions in the repository but that also didn't work for me
11:32 < zoq> Did you start with a clean build, so, removed the build folder?
11:32 -!- temp [7c42adce@] has joined #mlpack
11:33 < hemal[m]> no, i'll try that now
11:33 -!- temp is now known as heisenbuug
11:35 < Manav-KumarGitte> Yes
11:39 < hemal[m]> clean build seems to work, will take a while to finish ...
11:39 < hemal[m]> I wanted to save time, but now we wait ... 😂
11:40 < jeffin143[m]> hemal ( that sometimes happen
11:41 < jeffin143[m]> Hen pre compile header
11:41 < jeffin143[m]> Is changed
11:41 < jeffin143[m]> Run make clean
11:42 < hemal[m]> what is the hen precompiled header, what is it used for ?
11:42 < heisenbuug> Hey mentors,
11:43 < heisenbuug> I have submitted my draft proposal for the Visualization Tool Projcet, hope I can get some feed back regarding the prposal.
11:43 < heisenbuug> Thank you.
11:43 < heisenbuug> Project*
11:44 < heisenbuug> proposal*
11:47 < zoq> hemal[m]: Used to accelerate the build process.
11:55 < hemal[m]> zoq: got it, thanks!
11:55 < jeffin143[m]> rcurtin : should we record zoom meeting and put that up , so that whomsoever misses it can revist it and know what was there
11:55 < jeffin143[m]> :) no documentation needed , zoom does give you the option to record the video
11:56 < jeffin143[m]> hemal (
11:56 < jeffin143[m]> Sorry its not hen precompiled header
11:56 < jeffin143[m]> It was when precompiled header*
11:58 < hemal[m]> jeffin143: oh, ok, thanks
11:59 < zoq> jeffin143[m]: We would have to ask everyone for permission, don't think that's feasible, we could write a protocol, but not sure we have time for that.
12:08 < PrinceGuptaGitte> I have submitted my proposal draft through GSOC registration. Looking forward to every suggestion for improvement.
12:09 < Saksham[m]> ryan for the draft do we need to answer all the personal and optional questions, or just the project proposal works for now ?
12:11 < zoq> Saksham[m]: For the draft that's fine, we don't provide feedback for personal questions :)
12:11 < Saksham[m]> thanks
12:30 < heisenbuug> zoq, who is mentoring the visualization tool project? It's listed as Ryan Birmingham, but I was not able to get in touch with him via icr or mailing list...
12:34 < zoq> heisenbuug: Some of us are currently adapting to a new situation, so we can't provide responses as fast as we liked to, so please be patient.
12:35 < heisenbuug> ohhh, sorry...didn't mean to rush you guys...with all this going on in the world right now everyone is disturbed.
12:36 < zoq> exactly
12:40 < heisenbuug> I am glad that we can contribute to opensource via sitting at home...
12:40 < heisenbuug> And GSoC is anyways a work from home thing...
12:41 < zoq> I think a lot of people use the equipment at their universities, for GSoC as well.
12:42 < heisenbuug> Yea, even my university is closed and we all can back from the colleges to our home.
12:43 < Param-29Gitter[m> Yea... some students are facing these issues
12:44 < heisenbuug> In india govt as issued an order to close all the universities.
12:46 < sreenik[m]> I never thought of using my uni computers, would have saved a ton of time spent in compiling on my laptop
12:47 < heisenbuug> Yea, hardware will a considerable issue this season...
12:48 < chopper_inbound4> I was staying at our hostel and using university's server but yesterday they told us to vacate hostels :(
12:51 < sreenik[m]> heisenbuug: It generally always is. Many gsoc-ers I have met have bought a pc after gsoc (including me), and almost all of us only had a crappy laptop at that time. But it never occured to me about using uni PCs :(
12:52 < heisenbuug> Even I dont have a pc setup, staying in hostels and setting up a pc is difficult...
12:53 < sreenik[m]> On the other hand I remember shifting from xfce to i3 (i3 is just a window manager, not a complete GUI) just to free up 200Mb of RAM, took a while to set up but I have never looked back since.
13:01 < heisenbuug> Ohhh, I run dual boot windows and Ubuntu
13:01 < heisenbuug> Windows for gaming only.
13:03 < favre49> Does anyone here use vim for development?
13:04 < favre49> It would be great if you would share your vimrc :)
13:06 -!- heisenbuug [7c42adce@] has quit [Ping timeout: 240 seconds]
13:24 < birm[m]1> I need to figure out why I'm not being notified of irc, especially mentions.
13:26 < birm[m]1> It looks like I've just missed heisenbuug
13:45 < Param-29Gitter[m> Hey @brim I solved the kernel issue #2315.
13:45 < Param-29Gitter[m> @birm *
13:46 < Param-29Gitter[m> Hey @birm I solved the kernel issue #2315.
13:54 < LakshyaOjhaGitte> Hi @favre49 are you talking about the vim editor ?
13:55 < LakshyaOjhaGitte> In Linux
13:56 < favre49> Yup
13:59 < chopper_inbound4> it seems cv is requested in the project proposal. Can we just put link to cv?
14:22 < zoq> chopper_inbound4: We don't require a CV, you can answer the basic question in a couple of words.
14:28 < kartikdutt18Gitt> Hey everyone, Did anyone try the new Github app?
14:29 < kartikdutt18Gitt> *mobile app
14:29 < himanshu_pathak[> kartikdutt18 (Gitter): Yeah it's nice to use it
14:30 < Param-29Gitter[m> Yea.
14:31 < PranavReddyP16Gi> @matrixbot I used this setup : plus some extra lines for preference :
14:31 < Param-29Gitter[m> now we do not have to use browser to check updates of our PR's.
14:33 < kartikdutt18Gitt> Yes, it's really nice.
14:41 -!- favre49 [~favre49@] has quit [Quit: leaving]
14:44 < chopper_inbound[> <kartikdutt18Gitt "Yes, it's really nice."> With dark mode :)
14:45 < chopper_inbound[> zoq: ok. Thanks :)
15:19 -!- togo [~togo@2a02:6d40:34db:fd01:dd4d:2a2:670c:c3c7] has joined #mlpack
15:24 < SriramSKGitter[m> Is there a word/page limit for the proposal? I'm afraid mine's getting a bit too long :)
15:54 < M_slack_18> Is there already a PR for proximal policy optimization
15:59 < chopper_inbound[> Adarsh: yes, refer to <>
16:00 < M_slack_18> Also rainbow is partially complete?
16:16 < chopper_inbound[> I don't see much of it, anyway <> might be useful.
16:19 < zoq> SriramSKGitter[m: There is no limit, but it would be great if you can provide the information in a compact form.
16:20 < kartikdutt18Gitt> Hey zoq, Would it be okay to include relevant internships and hackathons as part of the proposal?
16:21 < M_slack_18> chopper_inbound Ok thanks
16:21 < zoq> kartikdutt18Gitt: Sure, I think that is a great idea.
16:21 < kartikdutt18Gitt> Great. Thanks a lot.
16:22 < M_slack_18> Other than the papers mentioned on the idea page is there something else I could implement too
16:22 < M_slack_18> Specifically for rl
16:22 < SriramSKGitter[m> @zoq : Thanks, I'll keep that in mind.
16:24 < zoq> M_slack_18: Yes, the papers listed are just ideas.
16:24 < himanshu_pathak[> zoq: I want to know that if I am designing a class such as DBN can I include a simple code snippet for that to show the functionality
16:24 < himanshu_pathak[> in my proposal
16:25 < himanshu_pathak[> sorry *architecture
16:28 -!- heisenbuug [7c42adce@] has joined #mlpack
16:31 < zoq> himanshu_pathak[: Absolutely, pseudo code is often really helpful to support an idea.
16:33 < himanshu_pathak[> zoq: Thanks I will add it
16:45 -!- r-purohit7 [75f181ff@] has joined #mlpack
16:55 -!- r-purohit7 [75f181ff@] has quit [Remote host closed the connection]
17:48 < heisenbuug> We are having video meet-up, right?
17:50 < rcurtin> yeah, in about ten minutes I'll join in
17:51 < heisenbuug> yea, even I will...
17:51 < heisenbuug> This wil be my first time though
17:58 < sreenik[m]> My desk is so stocked with snacks that there's no place for my laptop. I'll join through my PC (no camera or mic :( )
18:03 < zoq> Can't make it.
18:06 < rcurtin> zoq: no worries, see you another time :)
18:13 < naruarjun[m]> How do I join??
18:14 < naruarjun[m]> Or is it not for everybody?
18:15 < SriramSKGitter[m> @naruarjun :
18:43 < Manav-KumarGitte> make[2]: *** No rule to make target '/usr/local/lib/libboost_program_options.dylib', needed by 'lib/libmlpack.3.2.dylib'. Stop.
18:43 < Manav-KumarGitte> make[1]: *** [CMakeFiles/Makefile2:442: src/mlpack/CMakeFiles/mlpack.dir/all] Error 2
18:43 < Manav-KumarGitte> make: *** [Makefile:141: all] Error 2
18:44 < rcurtin> grep -ri CMakeCache.txt
18:44 < rcurtin> that's wrong
18:44 < rcurtin> grep -ri boost CMakeCache.txt
18:44 < rcurtin> ^^ that's the right one
19:02 * Manav-KumarGitte sent a long message: < >
19:03 < rcurtin> Manav-KumarGitte: okay, it looks like CMake thinks that boost is available in /usr/local/lib/
19:03 < rcurtin> but you said that the file /usr/local/lib/libboost_program_options.dylib doesn't exist?
19:10 < rcurtin> kartikdutt18Gitt: I created this issue for tracking what needs to be done to the new models/ repo:
19:11 < rcurtin> I'll probably get through some of those today, but perhaps some of those warrant opening more in-depth issues
19:12 < pickle-rick[m]> has the old models repo been taken down?
19:13 < rcurtin> I just renamed it to examples/, everything is intact
19:14 < pickle-rick[m]> ah i see phew 😂
19:15 < rcurtin> the new models/ repo has the same git history
19:15 < rcurtin> (I just merged in the examples/ repo)
19:15 < rcurtin> so basically they are clones, but all the issues and PRs went to the examples/ repo
19:17 < kartikdutt18Gitt> Ahh, That's great. Thanks a lot. I will reopen some of my PRs, so that we can started with a models repo as well.
19:19 < rcurtin> sounds good; I won't be able to handle everything today probably, but I hope to at least get issues open so that we have written down the things that should be done
19:19 < sreenik[m]> Btw rcurtin we are still having a good chat in the room after you left, much unlike what generally happens :)
19:20 < rcurtin> sreenik[m]: awesome, I am stuck in a far less interesting situation :)
19:21 < kartikdutt18Gitt> Yes that makes a lot of sense. I think we already have people starting to taking some of the issues that I opened. Hopefully we could finish up with restructuring in the next couple of weeks.
19:24 < sreenik[m]> rcurtin: :)
19:31 -!- heisenbuug [7c42adce@] has left #mlpack []
21:27 -!- ImQ009 [~ImQ009@unaffiliated/imq009] has quit [Quit: Leaving]
22:10 < metahost> Ugh, missed the conference again. :'(
22:11 < metahost> I was working on the proposal and lost track of time. :))
23:23 < rcurtin> no worries, there will be more :)
23:49 -!- togo [~togo@2a02:6d40:34db:fd01:dd4d:2a2:670c:c3c7] has quit [Quit: Leaving]
--- Log closed Fri Mar 20 00:00:36 2020