07:49 < MayankRaj[m]> Hey everyone I was thinking of implementing some pooling methods like Fractional max pooling, mixed pooling, gated pooling and tree pooling as a GSoC project. Can this be a valid project idea for this year??

08:21 < RishabhGargGitte> Hey @rcurtin , actually the documentation on the website is up-to-date.

08:21 < RishabhGargGitte> What happened yesterday with me was that I google searched "mlpack bindings" and the first link that popped up was https://mlpack.org/doc/mlpack-3.1.0/doxygen/bindings.html

08:22 < RishabhGargGitte> Which is actually the documentation for version 3.1.0.

08:24 < RishabhGargGitte> The documentation of the stable release is updated. I guess we need to mention this in the documentation too that never google search for mlpack on google, it redirects to older versions xD

08:24 < RishabhGargGitte> (edited) ... for mlpack on google, it ... => ... for mlpack, it ...

09:41 < MayankRaj[m]> Hi, I'm Mayank Raj and i am really interested in applying for Gsoc in mlpack. I am planning of implementing the following :-

13:01 < zoq> MayankRaj[m]: Hello, yes this could be a good GSoC project.

14:36 < iamarchit123Gitt> Hi @zoq @rcut

14:41 < AakashkaushikGit> hey @zoq update on the [PR](https://github.com/zoq/mlpack/pull/1) only `brnn` class is left and there are only two tests which fail and those are in the feedforward and serialization tests. if you can take a look or give feedback that would be awesome, after which i can update the existing models in the model repo and even add a new one.

14:42 < AakashkaushikGit> Anyone else that would like to review that PR is really welcome, it is a bit long (really sorry for that)

14:46 < zoq> AakashkaushikGit: Thanks, I'll take a look later today.

14:47 < zoq> iamarchit123Gitt: Hello there.

15:12 < AakashkaushikGit> thanks @zoq ( ╹▽╹ ) (just trying out these emojis)

15:23 < RishabhGargGitte> Hello, can someone please explain what does `FitnessFunction::template` mean [here](https://github.com/mlpack/mlpack/blob/825bfe5dff116a892f825364dabf4a8cc1535812/src/mlpack/methods/decision_tree/best_binary_numeric_split_impl.hpp#L129). I did try to look online but couldn't find it because I don't know what is this concept called. I am not a very advanced C++ programmer, a small help will be highly appreciated.

15:23 < RishabhGargGitte> Thanks!

15:27 < rcurtin[m]> Rishabh Garg (Gitter): this is a bit of a rabbit hole, maybe start here? https://womble.decadent.org.uk/c++/template-faq.html#disambiguation

15:27 < rcurtin[m]> basically there are some instances in which `typename` is not enough to tell the compiler that it is a dependent templated type, so you also need to add `::template`

15:27 < rcurtin[m]> I would not say it is pretty... :)

15:30 < RishabhGargGitte> Thanks for the link, i'll read it. Just to clarify, generally we require this syntax when we have a template inside a template. Am I right ?

15:30 < rcurtin[m]> yeah, that is a decent summarization of when we need it

15:31 < rcurtin[m]> I think there might be some additional complexities on top of that, but that's typically the mental rule I use in my head to decide whether a `::template` is needed, and then the compiler tells me if I am right or wrong :)

15:32 < RishabhGargGitte> Okay, right now your words aren't making much sense to me, but hopefully they will once I go through that link. :)

15:32 < rcurtin[m]> 👍️ for what it's worth, templates can be extremely complicated... it took me many years to learn them "well", and there are always surprising new things I didn't know

15:33 < RishabhGargGitte> Also, we have made `EvaluatePtr` a static function. Does this also has something to do with this too ?

15:33 < rcurtin[m]> I don't think `static` should play into that, no, just the fact that it is a doubly-derived template type (sorry if that wording doesn't make much sense, maybe the link will help)

15:35 < RishabhGargGitte> I am having many questions in my mind, but I will hold them back for now and first read that resource. After that I will ask if I have any further questions :). Thanks again!

15:35 < rcurtin[m]> the entire C++ FAQ is great, although it can take a while to get through (and maybe a few readings to understand)

15:36 < jonpsy[m]> Hey, could someone verify if my mail (yet again) got detoured to spam?

15:37 < rcurtin[m]> jonpsy: this one? http://knife.lugatgt.org/pipermail/mlpack/2021-March/004469.html

15:38 < zoq> jonpsy[m]: It's in my spam folder.

15:38 < jonpsy[m]> Dang it

15:38 < rcurtin[m]> I truly wish I knew what I could do to get mlpack list emails to not end up in spam in GMail, but there is no way that Google provides to handle that

15:38 < jonpsy[m]> Okay, I think I realise the problem

15:38 < jonpsy[m]> It's blocking my college id

15:39 < rcurtin[m]> jonpsy: I doubt that the issue is on your end... your posts to the list are successful

15:39 < jonpsy[m]> I'll try usig my gmail and send the same mail. If you guys don't mind?

15:39 < zoq> Not at all.

15:39 < rcurtin[m]> feel free, but note that not everyone on this list is in this channel, so you might want to add a note to the top of your email indicating why you're resending

15:40 < jonpsy[m]> Will do, thanks

15:40 < rcurtin[m]> otherwise you might get a "please be patient and wait for a response" 😃

15:40 < jonpsy[m]> Hahah wouldn't want that

15:49 < jonpsy[m]> Boom! Solved it, looks like our mailing list invalidates my college id. Should I put in issue section? =P

15:53 < rcurtin[m]> that's not an issue with the list software, as I described, it is an issue with the Google mailservers

15:53 < rcurtin[m]> see if you can figure out how to get Google support to help... good luck...

15:53 < jonpsy[m]> My gmail seems to work, so i dont mind :)

16:04 < zoq> Hello everyone, mlpack video meet-up in about one hour, see https://www.mlpack.org/community.html for more details.

16:05 < brongulusGitter[> What usually happens in the meet-ups?

16:08 < zoq> brongulusGitter[: We don't really have an agenda, often just some random stuff, we also try to answer questions about mlpack issues.

16:08 < RishabhGargGitte> Hey @rcurtin, I went through that FAQs, and I have the following interpretation. Here, `EvaluatePtr` is a function of a class i.e. `FitnessFunction`. Now, since it is a template parameter, compiler doesn't know whether it refers to an object or a type, it creates ambiguity. Thus to resolve that ambiguity, we are using `FitnessFunction::template` to tell it that it is a class.

16:08 < RishabhGargGitte> Although, if `FitnessFunction` were a `type` then we would have instead used `typename FitnessFunction`. Did I get it right ?

16:12 < rcurtin[m]> I think that is right, but also note that `EvaluatePtr` is itself a templated function

16:13 < rcurtin[m]> (if `EvaluatePtr` did not take that `bool` template parameter, we would not need the `::template`)

16:13 < brongulusGitter[> Sounds fun, are passive attendees allowed? :)

16:17 < AakashkaushikGit> Haha totally yes !

16:26 < RishabhGargGitte> > `rcurtin (@ryan:ratml.org)` (if `EvaluatePtr` did not take that `bool` template parameter, we would not need the `::template`)

16:26 < RishabhGargGitte> Ahh, yes. Now I understand it. Thanks a lot for explaining in such detail. I really appreciate your time.

16:26 < rcurtin[m]> 👍️

17:35 < CssTejaGitter[m]> Hello, i am trying to build mlpack from source in windows, i get this error, Could NOT find cereal (missing: CEREAL_INCLUDE_DIR) (Required is at least

17:35 < CssTejaGitter[m]> version "1.1.2") tried including cereal folder in the deps, still the same issue.

17:54 < jonpsy[m]> zoq hi would you mind reviewing the ZDT test suite in near future?

18:00 < rcurtin[m]> CssTeja (Gitter): I think `CEREAL_INCLUDE_DIR` should be the directory containing the `cereal/` directory; so, if you have `cereal.hpp` in, say, `C:/cereal/cereal.hpp`, then actually you would want `CEREAL_INCLUDE_DIR="C:/"`