00:03 < zoq> Yeah, so after the modification of the SGD class, we haven't updated all the necessary function.

00:03 < zoq> Take a look at line 94 in ffn.hpp this should look similar to line 143 in logistic_regression.hpp.

00:04 < zoq> I'll can make the changes, if you like, probably not before tomorrow.

00:07 < kris1> Oh i see. I will try to fix it if i can't then i will let you know.

00:08 < kris1> Maybe i think then i should write the test based upon logistic regression first.

11:10 < pvskand> When you run all the tests, is there any error found?

12:17 < zoq> pvskand: Since some tests are probabilistic, it's possible that some tests fail e.g. they don't converge in the specified number of iterations.

12:20 < pvskand> Okay. Well actually I submitted a PR on issue 722 and the Travis failed

12:21 < pvskand> I am guessing this is the reason

12:21 < pvskand> https://github.com/mlpack/mlpack/pull/946

12:21 < pvskand> Here is the link

12:22 < zoq> yes, this is an error that was fixed, yesterday.

12:24 < pvskand> So doesn't Travis build again automatically?

12:29 < zoq> No, Travis rebuilds the PR once you pushed something. I can restart the build if you like, but we know that the error you see isn't related to your changes.

12:30 < pvskand> Okay. Also one more question, can we run Travis locally? Or is it equivalent to running the tests using bin/mlpack test

12:32 < zoq> yeah, basically what Travis does is running 'cmake ..', 'make' and 'mlpack_test': https://github.com/mlpack/mlpack/blob/master/.travis.yml that's the Travis config that includes all steps

12:33 < pvskand> Okay thanks.

16:53 < sagarbhathwar> Hi, there . I am getting an error while compiling mlpack

16:53 < sagarbhathwar> Example : /usr/bin/ld: /usr/lib/libbfd.a(format.o): relocation R_X86_64_32S against symbol `binary_vec' can not be used when making a shared object; recompile with -fPIC

16:53 < sagarbhathwar> And many such errors

16:53 < sagarbhathwar> OS - Arch Linux, C++ compiler - gcc-6.3

16:54 < sagarbhathwar> cmake flags I am using is DEBUG=ON

16:54 < sagarbhathwar> I'm not able to figure out the source of this error

16:57 < sagarbhathwar> Oh, now I removed DEBUG=ON cmake flag and compilation proceeds without errors

17:06 < shikhar> Probably something to do with the backtrace printing mechanism in mlpack

17:06 < sagarbhathwar> Did I have to compile armadillo with debug on?

17:09 < shikhar> If you need the debug output, error checking, backtraces etc.

17:55 < pengpeng> Hi, I just got around to building mlpack, but I am having some trouble with the make command. The error I am getting is fatal error: /usr/include/clapack.h: No such file or directory #include ARMA_INCFILE_WRAP(ARMA_CLAPACK) - can anyone lend a hand? :)

17:55 < pengpeng> I have tried installing lapack and the C interface for lapack, but it still gives me this error

20:16 < travis-ci> mlpack/mlpack#2092 (master - 8b671ea : Ryan Curtin): The build was broken.

20:16 < travis-ci> Change view : https://github.com/mlpack/mlpack/compare/36470e9f6d3b...8b671eaf0612

20:16 < travis-ci> Build details : https://travis-ci.org/mlpack/mlpack/builds/212765168