00:40 < rcurtin> unclear why that build failed... I'll run it again

04:07 < mulx10> zoq : However If I increase the size of sequential model, the size of .bin file does not increase, but the total size of model increases. (https://pastebin.com/Yk5X7VU6)

04:11 < mulx10> zoq: Since it's a pointer I am also saving via * operator. Even that does not help.

04:11 < mulx10> Thanks

04:26 < jenkins-mlpack2> Project docker mlpack nightly build build #323: FAILURE in 12 min: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/323/

04:40 < Suryo> zoq: about the unconstrained optimization bit - yes, I totally agree. I'm currently documenting functions to try out for testing, though most of them are standard test cases. I'm making a PDF document and I'll share that with you soon and explain my plan to you.

04:40 < Suryo> zoq: Regarding your comments on my PR - I've seen them, but I will address them shortly. Right now, I want to be sure of the functions we will test.

04:44 < Suryo> zoq: And regarding https://bbcomp.ini.rub.de/ this seems really great! My only concern is that the deadline seems really really close. 30th June.

04:45 < Suryo> I'd be really excited to participate, though. I mean, PSO is good in this respect: we don't need gradient or functional information. But to make it battle-ready, we should definitely think of something on top of plain PSO. Maybe something like parallel swarms?

04:45 < Suryo> That's just an idea.

05:39 < jeffin143> rcurtin : does mlpack documentation in website gets updated automatically

05:40 < jeffin143> Bcse it's still the old one even after changing the inconsistency in adaboost

06:34 < jeffin143> Zoq , rcurtin : also a doubt , suppose if I have a file names as temp.txt in source code , but when I build the file and install it , the text file won't be there at install location. If you could tell me how to edit cmake so that temp.txt file is installed along with the build

07:06 < mulx10> jeffin143 : you can copy the file as in

07:07 < mulx10> https://github.com/mlpack/models/blob/627e39b1b90a5c80c2ee56295087912f6430dc88/vae/vae_mnist/CMakeLists.txt#L46

07:07 < jeffin143> Mulx10 , wrote this , but doesn't seem to work

07:08 < jeffin143> Do we have to run make again* or do we have to start over from cmake ../. ?

07:08 < mulx10> from cmake ..

07:08 < mulx10> bcoz u changed cmakelist

07:09 < jeffin143> Ahhaa I see *

07:09 < jeffin143> Thanks a ton

07:09 < jeffin143> Would do so , and let you know wether it worked

07:09 < mulx10> forgot to add the handle jeffin143, sorry

07:09 < mulx10> jeffin143: sure

07:09 < jeffin143> No issues with it 😊

07:56 < jeffin143> Mulx10 : it worked thanks

08:51 < jenkins-mlpack2> Project docker mlpack weekly build build #47: STILL FAILING in 8 hr 12 min: http://ci.mlpack.org/job/docker%20mlpack%20weekly%20build/47/

09:34 < zoq> mulx10: I see, let's replicate the serialization of the FFN class including the swap and assign operator.

09:38 < zoq> Suryo: Sounds good, about bbcomp I just referenced it as another possibility to test the implementation, but we can definitely submit something, probably for the next round?

12:41 < chandramouli_r> zoq: I have some problem with the test cases will building with ci-travis. Can you just have a look

12:41 < chandramouli_r> https://travis-ci.org/mlpack/mlpack/builds/531364961?utm_source=github_status&utm_medium=notification

13:09 < zoq> chandramouli_r: Can you set mlpack::Log::Info.ignoreInput = true; and mlpack::Log::Warn.ignoreInput = true; before the test starts, so that we will see the complete output, also does this pass on your local machine?

14:02 < rcurtin> jeffin143: the website gets rebuilt using doxygen documentation nightly

14:03 < rcurtin> it seems to be up to date for mlpack-git: http://mlpack.org/doc/mlpack-git/cli_documentation.html#adaboost

14:08 < jeffin143> rcurtin : forgot to change to mlpack-git , yeah its updated :), sorry

14:21 < rcurtin> jeffin143: no worries :)

14:46 < jeffin143> Mulx10 : suppose I am opening a file in temp.hpp ( temp.txt) which is installed in include/mlpack/* directory , through cmake just as we discussed in afternoon * , so everything is ok

14:47 < jeffin143> But , suppose now I make a program where I #include <mlpack/core/temp.hpp>

14:47 < jeffin143> And then just run it , the temp.hpp doesn't open the file because, it searches for temp.txt in the current directory where I am running the new program

14:48 < jeffin143> But temp.txt is saved in include/mlpack/* directory

14:48 < jeffin143> Umm anyway to get over it ..??

14:56 < jeffin143> Also just to make things clear * , suppose if I include whole path in temp.hpp ( as in /use/local/include/mlpack/core/temp.txt)

14:57 < jeffin143> But this is not perfect since a user may install mlpack anywhere in his computer ( both temp.hpp and temp.txt) are in same directory

14:58 < jeffin143> If whole path is given , it works perfectly alright* , but we need some other way out

15:00 < mulx10> jeffin143: there are options like CMAKE_CURRENT_SOURCE_DIR, PROJECT_BINARY_DIR from where u can copy temp.txt to the required place

15:01 < mulx10> zoq: I did not understand, did you mean copy ffn swap, assignment func to sequential?

15:43 < ShikharJ> saksham189: I reviewed both of your PRs, and they seem to be good to go. I'm currently scheduling a job on the savannah for them, and if the output checks out, I'll merge them at the earliest. If you wish, we can discuss and start off with the rest of the planned goals in your proposal.

15:44 < ShikharJ> Toshal: I'll try review your PRs shortly.

15:44 < zoq> mulx10: right

16:11 < Toshal> ShikharJ: No, problem.

16:12 < Toshal> I will complete WGAN implementation with Dual Optimizer till tommorow.

16:17 < ShikharJ> Toshal: That's fast, feel free to take it slow for now :)

16:29 < Toshal> ShikharJ: Actually, I would get busy with my exams soon so thinking to finish it off quickly.

16:39 < zoq> jeffin143: Couldn't you use a hpp file and treat it as a txt file?

16:50 < jeffin143> Zoq : I ll try that out, and let you know if that works

21:55 < favre49> zoq: I'll continue our discussion in a couple days, I hope that's fine. I'm pretty busy shifting hostels and stuff, but I'll be back home on 15th and have the time to brainstorm some

23:01 < zoq> favre49: Sounds good.