07:39 < SachaD> Thank you, zoq

08:10 < jenkins-mlpack2> Yippee, build fixed!

08:10 < jenkins-mlpack2> Project docker mlpack nightly build build #585: FIXED in 2 hr 56 min: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/585/

08:55 < SachaD> I corrected the input file https://yadi.sk/d/RxhD2xALODmy2w

08:55 < SachaD> then i write the code: in .h file: arma::mat dataset, trainData, outLabels; mlpack::ann::FFN<> model;

08:56 < SachaD> in .cpp file: mlpack::data::Load(csvFilePath.toStdString(), dataset, true);

08:56 < SachaD> trainData = dataset.submat(0, 0, dataset.n_rows - 1, dataset.n_cols - 1);

08:56 < SachaD> outLabels = dataset.submat(0, 0, dataset.n_rows - 1, dataset.n_cols - 1);

08:57 < SachaD> outLabels.for_each( [](arma::mat::elem_type& val) { val += 1; } ); //I read that output label can not be 0.

08:58 < SachaD> model.Add< mlpack::ann::Linear<> >(31, 31);

08:58 < SachaD> model.Add< mlpack::ann::SigmoidLayer<> >();

08:58 < SachaD> model.Add< mlpack::ann::Linear<> >(31, 31);

08:58 < SachaD> model.Add< mlpack::ann::LogSoftMax<> >();

08:59 < SachaD> model.Train(trainData, outLabels);

09:00 < SachaD> I got error: error: Mat::operator(): index out of bounds terminate called after throwing an instance of 'std::logic_error'

09:00 < SachaD> What I did wrong?

11:39 < zoq> SachaD: Can you create a minimal example to reporduce the issue?

12:16 < Soham> i want to install mlpack for python but there is some problem(Windows User)! As per the documentation it's straight and simple using either "pip install mlpack/mlpack3" or using "conda install -c mlpack mlpack" but neither seem to work!!any help will be highly appreciated!Thanks!

14:28 < HimanshuPathakGi> Soham : Can you please share error message for windows build you can refer to this https://www.mlpack.org/doc/mlpack-git/doxygen/build_windows.html

14:30 < HimanshuPathakGi> (edited) ... message for ... => ... message so that we can understand what the problem is for ...

15:42 < kumaran[m]> <zoq "kumaran: Hello, there is a PR op"> PR number?

18:38 < metahost> Soham: if you could share the error log (via pastebin.com) then it would be great!

18:43 < rcurtin> Soham: metahost: I am still working out getting the Windows Python bindings to build so I can upload them to pip. I think conda should work though?

18:43 < rcurtin> might be a good idea to file an issue with the conda Github repository if not

18:46 < zoq> conda-forge has linux builds only, but you are right we should probably just open a PR to enable the windows build

18:47 < zoq> kumaran[m]: https://github.com/mlpack/ensmallen/pull/152 this is the one I was talking about

18:53 < rcurtin> zoq: oh, really? weird, I thought I worked with the conda-forge maintainers to fix that

18:54 < rcurtin> I don't know conda at all, so I'm not sure what the right way to go is

18:54 < rcurtin> if you want I can ping the conda mlpack maintainer, or, maybe we can open an issue on mlpack's github and someone interested in contributing might give it a shot?

19:04 < zoq> Just looked at: https://github.com/conda-forge/mlpack-feedstock and https://anaconda.org/conda-forge/mlpack/files

19:05 < zoq> I can open an issue on the mlpack github, there is probably someone who is interested in take a closer look into the issue.

19:10 < rcurtin> sounds good, I think that might get some interest (hopefully :))

19:51 < HimanshuPathakGi> I think I would like to work on this issue

19:52 < HimanshuPathakGi> (edited) I think Iwould ... => I would ...

20:02 < HimanshuPathakGi> So, here we need to enable windows build. zoq can you please provide more context

20:04 < zoq> HimanshuPath: Right, so as you can see on https://github.com/conda-forge/mlpack-feedstock the windows build is disabled, not sure what the reason is, so I guess you could checkout some other packges in the conda-forge repo that has the windows build enabled and see if you can replicate that for mlpack.

20:05 < zoq> They skip windows and macOS: https://github.com/conda-forge/mlpack-feedstock/blob/master/recipe/meta.yaml#L13

20:06 < HimanshuPathakGi> Ok I will try to find out

20:10 < zoq> Great, thanks!