05:29 < Toshal> ShikharJ: Sorry I missed the meeting.

05:29 < Toshal> sakshamB: I also liked your idea

05:30 < Toshal> So let me know if you are going to implement it.

05:30 < Toshal> Or should I work on it?

05:31 < Toshal> Just let me know your thoughts regarding the same.

05:33 < Toshal> Sorry for such a late response.

07:11 < rcurtin> ok, I'm at the airport for my flight home... time to review some PRs and finally respond to a lot of emails

07:11 < rcurtin> good thing it is a long flight :) (hopefully the internet works...)

07:39 < jenkins-mlpack2> Project docker mlpack nightly build build #378: STILL UNSTABLE in 3 hr 25 min: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/378/

11:47 < jenkins-mlpack2> Project mlpack - git commit test build #196: UNSTABLE in 45 min: http://ci.mlpack.org/job/mlpack%20-%20git%20commit%20test/196/

12:03 < travis-ci> mlpack/mlpack#7477 (master - f0976bd : Marcus Edel): The build is still failing.

12:03 < travis-ci> Change view : https://github.com/mlpack/mlpack/compare/c62088ad40c4...f0976bd3adbf

12:03 < travis-ci> Build details : https://travis-ci.org/mlpack/mlpack/builds/554992135

13:03 < sakshamB> Toshal: I’ll work on implementing the Padding layer and weight dimension visitors in a new PR. Just let me know if you could use the visitors for weight norm as well.

13:04 < favre49> zoq It looks like the Go version of NEAT, as well as the original has an option to specify the starting point. I think we could implement this using the serialization feature, what do you think?

13:08 < favre49> An issue I see is that the serialization currently serializes both the directed graph and the connection gene list, ideally it could build one from the other on loading. Is there something I could do about that?

13:08 < zoq> favre49: hm, or the task has some function that returns the starting point, similar to: https://github.com/mlpack/ensmallen/blob/master/include/ensmallen_bits/problems/easom_function.hpp#L56

13:09 < zoq> favre49: You could check if you are currently loading: https://github.com/mlpack/mlpack/blob/2326bf8bd0b380c17eed6b5b1c34fbfab1563774/src/mlpack/methods/hoeffding_trees/hoeffding_tree_model.hpp#L172

13:11 < favre49> Oh right, that work too. It's not clear to me which is better, though

13:13 < zoq> favre49: What about we go with both? If Nothing is specified in the config file, we go with what is specefied in the task.

13:14 < favre49> Sounds good, I'll do that. Hopefully that could fix the issues with the tests

13:15 < zoq> Yeah, let's see, I'm still debugging the code.

13:16 < favre49> I'll try to get it done by Monday, I have to travel a lot tomorrow unfortunately

13:16 < favre49> Thanks for the help!

13:29 < zoq> favre49: Sounds good, have fun.

14:00 < Toshal> sakshamB: I would say I need adding one or two visitor for getting the bias term and weight term separately. If you are going to implement it then it would be helpful for me for sure.

14:04 < jenkins-mlpack2> Yippee, build fixed!

14:04 < jenkins-mlpack2> Project mlpack - git commit test build #197: FIXED in 45 min: http://ci.mlpack.org/job/mlpack%20-%20git%20commit%20test/197/

14:09 < sakshamB> Toshal: if you get the dimension of the weight term then you can separate the bias since bias is stored at the end.

14:20 < travis-ci> mlpack/mlpack#7478 (master - ebaea0f : Ryan Curtin): The build was fixed.

14:20 < travis-ci> Change view : https://github.com/mlpack/mlpack/compare/f0976bd3adbf...ebaea0fdf69c

14:20 < travis-ci> Build details : https://travis-ci.org/mlpack/mlpack/builds/555019978