04:27 < tesseract28> Hello people

04:28 < tesseract28> I am trying to build mlpack on Mac using a conda environment

04:29 < tesseract28> Having trouble trying to link conda armadillo installation to cmake config

04:29 < tesseract28> CMake Error at CMake/FindArmadillo.cmake:326 (message): ARMADILLO_INCLUDE_DIR-NOTFOUND/armadillo_bits/config.hpp not found! Cannot determine what to link against.

04:29 < tesseract28> ^ any help would bee appreciated!

04:51 < rcurtin> tesseract28: when you are configuring mlpack with CMake, try setting ARMADILLO_INCLUDE_DIR and ARMADILLO_LIBRARY

04:51 < rcurtin> i.e.

04:51 < rcurtin> $ cmake -DARMADILLO_INCLUDE_DIR=/path/to/armadillo/include/ -DARMADILLO_LIBRARY=/path/to/libarmadillo.so ../

04:52 < rcurtin> (where you put whatever the appropriate paths are in place of the placeholders I wrote, of course :))

04:58 < rcurtin> KimSangYeon-DGU: I think I figured out the IRC log issue! the GNU 'date' command changed default behavior to print am/pm instead of 24 hour times and this appears to have caused the problem

04:58 < rcurtin> I'll see if my fix works tomorrow :)

04:59 < picklerick> hello rcurtin i was seeing through the open issue#1565 and i read through the comments and following that up i looked up issue#1582 and read kmeans_main.cpp should i follow up the steps that are done in #1582

04:59 < picklerick> to see where the inconsistency lies or error is coming . ?

05:03 < rcurtin> picklerick: I think #1616 would be a much better PR to look at for an idea. I left very many comments in #1582 because it changed a lot of things that did not make sense to change

05:03 < rcurtin> #1616 only does logistic_regression but there are some others that you could consider also

05:03 < rcurtin> thanks for looking into it! I am headed to bed now, back later

05:16 < tesseract28> rcurtin thanks! atleast the cmake is configured, onto build now!

05:26 < tesseract28> So the build has failed with this error

05:26 < tesseract28> In file included from /Users/tesseract28/anaconda3/pkgs/armadillo-9.200.4-blas_openblash25ba8c5_1203/include/armadillo:88:0, from /Users/tesseract28/Documents/mlpack-3.0.4/src/mlpack/core/arma_extend/arma_extend.hpp:52, from /Users/tesseract28/Documents/mlpack-3.0.4/src/mlpack/prereqs.hpp:108, from /Users/tesseract28/Documents/mlpack-3.0.4/build/src/mlpack/cotire/mlpack_CXX_prefix

05:29 < tesseract28> Users/tesseract28/anaconda3/pkgs/armadillo-9.200.4-blas_openblash25ba8c5_1203/include/armadillo_bits/include_superlu.hpp:95:53: fatal error: /opt/anaconda1anaconda2anaconda3/include/supermatrix.h: No such file or directory #include ARMA_INCFILE_WRAP(ARMA_SLU_SUPERMATRIX_H) ^ compilation terminated.

05:29 < tesseract28> Not quite sure how the wierd path '/opt/anaconda1anaconda2anaconda3/include/supermatrix.h' got generated

05:30 < KimSangYeon-DGU> My Internet is disconnected.

05:31 < KimSangYeon-DGU> rcurtin: Wow!! It' so glad for you to find the possible solution. :)

05:32 < KimSangYeon-DGU> So respectful!!

05:33 < ShikharJ> rcurtin: Do we still have a test machine like savannah online? I have a long weekend coming up, and among the things to get off in my to-do list is to finish up on CycleGAN PR, for which just some trial running and tweaking is remaining. I'll update the Ideas list then also, and take a look at places which are missing some love :)

05:36 < KimSangYeon-DGU> *I'm so glad

05:37 < KimSangYeon-DGU> :)

07:48 < picklerick> In file included from /home/whoopitydoop/mlpack-3.0.3/src/mlpack/tests/main_tests/logistic_regression_test.cpp:19:0:

07:49 < picklerick> /home/whoopitydoop/mlpack-3.0.3/src/mlpack/methods/logistic_regression/logistic_regression_main.cpp:18:10: fatal error: ensmallen.hpp: No such file or directory

07:49 < picklerick> #include <ensmallen.hpp>

07:49 < picklerick> ^~~~~~~~~~~~~~~

07:49 < picklerick> guys this is the error i am getting does this means i have to install the header of ensmallen and if so then how ??

08:01 < KimSangYeon-DGU> picklerick: You can build the ensmallen from the source code

08:01 < KimSangYeon-DGU> please refer to https://github.com/mlpack/ensmallen

08:02 < KimSangYeon-DGU> clone the repository

08:02 < KimSangYeon-DGU> cd ensmallen && mkdir build

08:02 < KimSangYeon-DGU> cmake ../

08:02 < KimSangYeon-DGU> make

08:02 < KimSangYeon-DGU> sudo make install

08:12 < KimSangYeon-DGU> Ah!! before doing cmake, you should move to 'build' directory. :)

08:13 < KimSangYeon-DGU> clone the ensmallen repository

08:13 < KimSangYeon-DGU> cd ensmallen && mkdir build

08:13 < KimSangYeon-DGU> cd build

08:13 < KimSangYeon-DGU> cmake ../

08:13 < KimSangYeon-DGU> make

08:13 < KimSangYeon-DGU> sudo make install

08:14 < KimSangYeon-DGU> If you want to use parallel execution when 'make' process, please use '-j' option. For example, make -j$(nproc)

08:16 < KimSangYeon-DGU> $(nproc) is to print the number of processing units available in your computer. :)

08:41 < jenkins-mlpack2> Yippee, build fixed!

08:42 < jenkins-mlpack2> Project docker mlpack nightly build build #204: FIXED in 3 hr 27 min: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/204/

10:29 < Ishmeher> Hi

10:36 < Ishmeher> I am relatively new to open-source and would want to work on the idea of implementing the Deep Learning Modules. Can someone suggest me where to start from?

11:54 < zoq> Ishmeher: Hello there, http://www.mlpack.org/involved.html and https://www.mlpack.org/gsoc.html should be helpful here.

13:33 < picklerick> In file included from /home/whoopitydoop/mlpack-3.0.3/src/mlpack/tests/main_tests/kmeans_test.cpp:20:0:

13:33 < picklerick> /home/whoopitydoop/mlpack-3.0.3/src/mlpack/methods/kmeans/kmeans_main.cpp:108:57: error: macro "PROGRAM_INFO" passed 10 arguments, but takes just 2

13:33 < picklerick> "@doxygen/classmlpack_1_1kmeans_1_1KMeans.html"));

13:33 < picklerick> ^

13:33 < picklerick> /home/whoopitydoop/mlpack-3.0.3/src/mlpack/methods/kmeans/kmeans_main.cpp:31:1: error: ‘PROGRAM_INFO’ does not name a type

13:33 < picklerick> PROGRAM_INFO("K-Means Clustering",

13:34 < picklerick> what could be the issue ??

13:50 < rcurtin> picklerick: try using mlpack 3.0.4, 3.0.3 is an older version

14:20 < rcurtin> ShikharJ: savannah is still online, feel free to use it

14:21 < rcurtin> tesseract28: I saw your messages from earlier, that is a strange path that got generated for sure

14:21 < tesseract28> Hello all, just re-iterating my issue. I am trying to build mlpack on Mac using a conda environment. The build fails with the following error.

14:21 < rcurtin> do you want to open an issue on Github? the output of cmake when you configure mlpack would be helpful too

14:21 < tesseract28> Hmm, yea

14:21 < rcurtin> I'm suspecting that the CMake file CMake/FindArmadillo.cmake is somehow generating weird paths

14:21 < tesseract28> Sure, will do right now.

14:21 < rcurtin> thanks

14:22 < rcurtin> I'm traveling tomorrow and over the weekend so I may not be able to look into it soon, but maybe someone else will have some ideas

14:23 < tesseract28> Sure! I ll try on a linux server in the meantime. But it would be nice to have to up and running on my own machine.

14:24 < rcurtin> definitely agreed

14:24 < rcurtin> I don't know how many people have built mlpack in a conda environment, so there may be a bug or two in the CMake configuration to work out

14:25 < rcurtin> CMake is a nice tool, but it's very hard to make it bulletproof :)

14:33 < tesseract28> Haha, I am not very well versed with using cmake. Hopefully will understand it better with this issue..

14:36 < rcurtin> tesseract28: just saw #1696, thanks for opening the issue

14:36 < tesseract28> So I have opened issue #1696.

14:38 < tesseract28> rcurtin: No problem! Thanks for the prompt responses.

14:38 < rcurtin> sure, happy to help out

20:14 < kassel> hey, will you participate in gsoc this year?

20:43 < rcurtin> kassel: our intention is to, yes

21:08 < picklerick> command for running the individual test suite for kmeans in mlpack

21:11 < picklerick> unknown location(0): fatal error: in "DecisionTreeMainTest/DecisionRegularisationTest": memory access violation at address: 0x00000000: no mapping at fault address

21:11 < picklerick> /home/whoopitydoop/mlpack-3.0.4/src/mlpack/tests/main_tests/decision_tree_test.cpp(202): last checkpoint: "DecisionRegularisationTest" test entry

21:12 < picklerick> *** 1 failure is detected in the test module "mlpackTest"

21:13 < zoq> picklerick: mlpack -t DecisionTreeMainTest/DecisionRegularisationTest

21:13 < zoq> picklerick: see https://github.com/mlpack/mlpack/pull/1674#issuecomment-455796577 for some additional infos

21:15 < picklerick> zoq so if i wanted to run test individually like for knn it is bin/mlpack-t knnTest how does this change for others?

21:16 < picklerick> got it thanks for guiding me

21:16 < zoq> great, just search for BOOST_AUTO_TEST_SUITE to get the test suite name and BOOST_AUTO_TEST_CASE for the test case.

21:18 < zoq> or bin/mlpack_test --list_content

21:18 < zoq> should list every test

21:27 < rcurtin> nice, I didn't know about --list_content

22:11 < picklerick> @rcurtin i did the changes in kmeans as were done in logistic regression test cases are running fine too how to check the bindings for renamed parameters and options

22:11 < picklerick> regarding issue #1565

23:24 < ShikharJ> rcurtin: Yeah, about that. I made a rather stupid blunder. I formatted my system, without saving the login credentials and the ssh key :/

23:27 < zoq> ShikharJ: I can add your ssh key if that helps.

23:38 < rcurtin> picklerick: but it doesn't make any sense to make this change for k-means, because the output of k-means isn't predictions

23:38 < rcurtin> we should only make this change for classifiers