mlpack IRC logs, 2020-01-22

Logs for the day 2020-01-22 (starts at 0:00 UTC) are shown below.

January 2020
--- Log opened Wed Jan 22 00:00:28 2020
00:41 -!- kyrre [uid207859@fsf/member/kyrre] has quit [Quit: Connection closed for inactivity]
00:45 -!- ibtihaj [cbd7be0c@] has quit [Ping timeout: 260 seconds]
02:33 -!- vancao [01343ad2@] has joined #mlpack
02:37 < vancao> Hi everyone. My name is Van from Viet Nam. I am currently a third year Computer Science student. I have experience in C++ and general knowledge of Machine Learning. Do you have any suggestions for a beginner like me to participate GSOC?.
03:47 -!- vancao [01343ad2@] has quit [Remote host closed the connection]
04:41 -!- M_slack_mlpack_U [slackml_15@gateway/shell/] has joined #mlpack
05:11 -!- vancao [01343ad2@] has joined #mlpack
05:12 -!- vancao [01343ad2@] has quit [Remote host closed the connection]
06:31 < metahost> vancao: Hi Van! You may check out and to get started. :)
08:44 -!- saumya [0e8bdfaf@] has joined #mlpack
08:50 -!- saumya [0e8bdfaf@] has quit [Remote host closed the connection]
11:48 -!- rajs123 [ca838902@] has joined #mlpack
11:55 -!- rajs123 [ca838902@] has quit [Remote host closed the connection]
14:06 < hunting-party104> Hi , I'm trying to build only the ann module from source but it keeps telling me `make: *** No rule to make target 'mlpack_ann'. Stop.`
14:07 < hunting-party104> * Hi , I'm trying to build only the ann module from source but it keeps telling me `make: *** No rule to make target 'mlpack_ann'. Stop.`
14:07 < hunting-party104> I can build a few modules like det, cf etc but for some like ann, adaboost, amf it gives me that error
14:12 < Nakul[m]> hey you can buid any particular module by getting the corresponding target . for ann you can build by navigate to it's make file for example
14:12 < Nakul[m]> ```/home/knakul853/Desktop/mlpack/mlpack/build/src/mlpack/methods/ann``` them running only make would work
14:12 < Nakul[m]> *then
14:13 < hunting-party104> So i can just go to the ann directory and type `make` instead of doing it from the build directory ?
14:17 < hunting-party104> Nope on typing make it doesnt do anything
14:19 * Nakul[m] sent a long message: < >
14:19 < Nakul[m]> hope it helps
14:20 < hunting-party104> Ah wait my bad i didnt see that i had to go into the build directory then the ann, I drectly went to ann the normal route
14:23 < zoq> You can only build targets/executables since there is no executable for the ann code, it will fail.
14:23 < hunting-party104> so is there no way to only build the ann module?
14:25 < zoq> right, you can disable the executables and build the lib only, to speedup the build.
14:26 < zoq> By using -DBUILD_CLI_EXECUTABLES=OFF in the cmake step.
14:27 < rcurtin> Nakul[m]: I would strongly advise against navigating the build directories and using make in the subdirectories
14:28 < rcurtin> instead use 'make <target>' from the top-level build directory
14:28 < rcurtin> please read the CMake documentation to understand more
14:30 < hunting-party104> Sorry im a bit new to cmake zoq how do i build only the lib ? By just typing `make` or `make lib`?
14:31 < hunting-party104> well `make lib` doesnt work so would `make` only build the libraries ?
14:31 < rcurtin> hunting-party104: the target is called 'mlpack' just to build so, 'make mlpack'
14:32 < rcurtin> if you're new to CMake, go read some resources about how to use CMake so that it makes sense---that will be a lot more productive in the long run :)
14:32 < Nakul[m]> > Nakul: I would strongly advise against navigating the build directories and using make in the subdirectories
14:32 < Nakul[m]> thanks i wasn't aware about that
14:33 < rcurtin> sure; it really makes a big difference to spend some time getting familiar with the tools, so even if it feels unproductive to read tutorials, etc., it actually is productive, and when you return to the code you will find yourself better positioned to make changes :)
14:33 < hunting-party104> ah ic rcurtin are there any good references for cmake or just a google search would do ?
14:37 < rcurtin> I don't have a preferred resource; I'd suggest searching google
14:39 < hunting-party104> rcurtin:
14:40 < hunting-party104> Also rcurtin i wanted to pursue issue #1631 which was opened by @akhandait whos not on this channel i guess, Could you give any pointers on how i could proceed ?
14:43 < rcurtin> the issue description seems pretty comprehensive to me. if not all of the description makes sense, spend time in the codebase learning it until the issue description fully makes sense
14:43 < rcurtin> then, open a PR with the solution
14:43 < rcurtin> or, I shouldn't say "the solution", just "a solution". there may be more than one :)
14:47 < hunting-party104> <rcurtin "the issue description seems pret"> yes the problem seems straight forward but im not able to understand a few things from the code , like for eg Sequential which calls the destructor only when `const bool model` is `false`
14:48 < hunting-party104> from the destructor
15:00 -!- ImQ009 [~ImQ009@unaffiliated/imq009] has joined #mlpack
15:07 < rcurtin> hunting-party104: my suggestion would be to spend more time with the Sequential layer and look at how it's used in hte tests, etc., to understand what the 'model' parameter is
15:10 < hunting-party104> Alright will do
15:53 < HimanshuPathakGi> Hey zoq, rcurtin windows build for ensmallen is now enabled all thanks to Marcelo now I will work for mlpack now
16:22 -!- angaria3 [a995c869@] has joined #mlpack
16:27 -!- angaria3 [a995c869@] has quit [Remote host closed the connection]
17:05 -!- Ayush [67763204@] has joined #mlpack
17:42 < hemal[m]> Are we having a video Meetup at 22nd, at 2200 UTC?
17:42 < hemal[m]> I would like to join this one. (Missed the last one ☹️)
17:42 < hemal[m]> Also, any points on the agenda of the Meetup ?
17:42 < abernauer[m]> Yeah that is the time.
17:43 < hemal[m]> Read the full email just now,
17:43 < hemal[m]> No agenda, got it, hope to see you all there 😊
17:44 -!- Ayush75 [67763204@] has joined #mlpack
17:46 -!- Ayush75 [67763204@] has quit [Remote host closed the connection]
17:53 < HimanshuPathakGi> How can we join video meetup
18:02 < HimanshuPathakGi> (edited) ... video meetup => ... video meetup?
18:02 < zoq> Just follow the zoom link in the mail.
18:02 < abernauer[m]> Set up a zoom account the link to the chat is in the mailing list btw.
18:04 -!- Ayush [67763204@] has quit [Ping timeout: 260 seconds]
18:05 < HimanshuPathakGi> ok thanks I will do
18:16 < rcurtin> can someone try to join the zoom for the weekly meetup? I think I have changed my zoom account settings so anyone should be able to join at any time
18:17 < rcurtin> so it shouldn't say "please wait for the host to start the meeting"... I hope :)
18:18 < zoq> Nope, "please wait for the host to start the meeting"
18:19 < abernauer[m]> Yeah I got the same thing.
18:20 < zoq> rcurtin: meeting id is 3820896170?
18:21 < birm[m]> I thought it was scheduled for 2200 UTC
18:22 < abernauer[m]> He is just confirming if it will start the call if he's not present
18:22 < abernauer[m]> Or meeting.
18:23 < rcurtin> zoq: yeah, that's the right id. I guess I didn't get it right
18:23 < rcurtin> let's see what happens if I actually schedule a meeting...
18:23 < rcurtin> actually I see, I just made the change in the wrong configuration page. can you try joining the 3820896170 meeting again?
18:27 -!- M_slack_mlpack_4 [slackml_16@gateway/shell/] has joined #mlpack
18:27 < zoq> looks good
18:27 -!- M_slack_mlpack_4 is now known as ToshalAgrawal[m]
18:28 < ToshalAgrawal[m]> Hi zoq just checking out slack
18:32 < ToshalAgrawal[m]> @freenode_rcurtin: I hope this works.
18:33 < ToshalAgrawal[m]> rcurtin: zoq: I will update the ideas page soon.
18:34 < zoq> ToshalAgrawal: Sounds good :)
18:42 * jeffin143[m] uploaded an image: (42KB) < >
18:42 < jeffin143[m]> Going to get my hands on one of this
18:42 < jeffin143[m]> Really awesome to wear contribution map :)
18:43 -!- HimanshuPathak[m [slackml_17@gateway/shell/] has joined #mlpack
18:43 < HimanshuPathak[m> Yeah It looks cool
18:45 < rcurtin> jeffin143[m]: nice! but do you have to get a new shirt every day? :)
18:45 < rcurtin> zoq: awesome, glad it works
18:46 < rcurtin> ToshalAgrawal[m]: let me know if you have any problems using slack... I think the bridge works fine and is stable though :)
18:46 < jeffin143[m]> :) They would be really heavy on my pocket , but i am really after swags
18:46 < jeffin143[m]> Probably have to mail for Mlpack stickers , Haven't done till now
18:57 < rcurtin> :) I'd be happy to toss some in the mail for you
19:01 -!- ImQ009 [~ImQ009@unaffiliated/imq009] has quit [Quit: Leaving]
19:52 -!- thePythonWizard [~thePython@] has joined #mlpack
20:04 -!- thePythonWizard is now known as sg-ai-vision
20:05 -!- sg-ai-vision [~thePython@] has left #mlpack ["Leaving"]
20:09 -!- sg-ai-vision [~sg-ai-vis@] has joined #mlpack
20:32 < Nakul[m]> can anyone help me with this ```difference{-nan%} between arma::sum(probabilities.col(i)){-nan} and 1.0{1} exceeds 1.0000000000000001e-05%```
20:45 -!- AbhinavvermaGitt [gitterab_5@gateway/shell/] has joined #mlpack
20:45 < AbhinavvermaGitt> Do mlpack can be built on the 32-bit window.
20:47 < zoq> Abhinavverma: yes
20:50 < rcurtin> Nakul[m]: do you understand the meaning of the error message?
20:51 < Nakul[m]> in probabilites matrix there are some value which are nan(not a number)
20:53 < Nakul[m]> and when we are passing to the boost for testing the this error messge shows up
20:53 < Nakul[m]> i tried checking by ```std::isnana(the value)```
20:54 < Nakul[m]> and then pass it for testing but seems not working..:(
21:06 < Nakul[m]> > Nakul: do you understand the meaning of the error message?
21:06 < Nakul[m]> not realy ?
21:08 < Nakul[m]> basically this error comes from here ``` BOOST_REQUIRE_CLOSE(arma::sum(probabilities.col(i)) , 1.0, 1e-5);
21:08 < Nakul[m]> ```
21:08 < Nakul[m]> * basically this error comes from here ``` BOOST_REQUIRE_CLOSE(arma::sum(probabilities.col(i)) , 1.0, 1e-5);```
21:11 < Nakul[m]> any refrence would be helpful , since simply googling seems not helpful
21:25 < rcurtin> ok, so a place to start would be to find documentation for BOOST_REQUIRE_CLOSE
21:25 < rcurtin> that's a part of the Boost Unit Test Framework
21:26 < rcurtin> once you understand that, you will be able to understand what has actually happened that has caused the test to fail
21:26 < rcurtin> then, you can read the rest of the test in order to understand what is being tested
21:26 < rcurtin> and from there you can reason to determine what is wrong
21:28 < Nakul[m]> thanks
21:43 < rcurtin> I know that I am not being immediately helpful by telling you what the issue is, but, if you take the time to dig in deeper and learn what's going on you will be much more effective overall :)
21:44 < rcurtin> also, unfortunately it looks like I won't be able to make the video meetup. I'm stuck in company meetings all afternoon
21:46 < Nakul[m]> > I know that I am not being immediately helpful by telling you what the issue is, but, if you take the time to dig in deeper and learn what's going on you will be much more effective overall :)
21:46 < Nakul[m]> These little replies and hints means a lot...
21:47 -!- sg-ai-vision [~sg-ai-vis@] has quit [Quit: Leaving]
21:54 < zoq> I wonder if this is a late "meeting" for everybody or just me :)
21:55 < rcurtin> well, at least someone who filled out the survey said it was a good time :)
21:57 < zoq> haha, maybe for someone it's lunch time.
22:01 < zoq> Will use the time to review a PR.
22:06 < rcurtin> I keep getting emails that say "your meeting participants are waiting", but did you manage to join ok?
22:32 < zoq> I did.
22:35 < zoq> Since we are 5 in the call, it worked.
22:47 < zoq> bin/mlpack_test -t QLearningTest
22:47 < zoq> bin/mlpack_test -t QLearningTest/CartPoleWithDQN
22:55 < KimSangYeon-DGU4> It was nice to have the meeting. Thanks, everyone!
22:56 < pickle-rick[m]> > bin/mlpack_test -t QLearningTest/CartPoleWithDQN
22:56 < pickle-rick[m]> Thanks. Will try this out!
23:04 < hemal[m]> Nice meeting you all, Thanks :)
23:05 < zoq> hemal[m]: Sorry haven't checked the chat, I couldn't hear you.
23:07 < hemal[m]> Oh, thanks for letting me know. It's fine. Thank you for the PR review too. :)
23:09 < zoq> hemal[m]: It's a nice addition, should be helpful for a lot of people.
23:28 < hemal[m]> zoq: Thanks
--- Log closed Thu Jan 23 00:00:29 2020