05:39 < RudraPatil[m]> Can we choose any dataset for Application of ANN Algorithms Implemented in mlpack ?

06:23 < Param-29Gitter[m> Hello, I wanted to know if anyone knows a data-set with 10M numeric samples for classification problem. I wanted it for bench-marking of Naive Bayes.

07:28 < bisakh[m]1> @<https://app.slack.com/services/BQJ5NGK9R|Param-29 (Gitter)>: Hey, you can check <https://datasetsearch.research.google.com/> for any public dataset.

15:15 < PranavReddyP16Gi> Looks like India is on lockdown for another three weeks hope everyone is staying safe

15:16 < Saksham[m]> Yeah anyone from Delhi ?

15:19 < PrinceGuptaGitte> Yes, now the whole India is in full lockdown. Also, I'm from Delhi.

15:20 < Saksham[m]> Stay Safe Guys !!

15:23 < LakshyaOjhaGitte> Yeah

15:24 < LakshyaOjhaGitte> 21 days of fate

15:27 < LakshyaOjhaGitte> anyone working on models repo?

15:30 < jeffin143[m]> Lot's of time for open source :)

15:32 < LakshyaOjhaGitte> XD

15:33 < LakshyaOjhaGitte> can be considered the silver lining

15:39 < kartikdutt18Gitt> Hey @ojhalakshya, I am.

15:42 < kartikdutt18Gitt> @ojhalakshya, if you want to work on the repo, there are some good first issue if you want to help out.

15:43 < kartikdutt18Gitt> You could take a look at [this](https://github.com/mlpack/examples/issues/66).

15:48 < LakshyaOjhaGitte> Thanks for the update on it :)

15:48 < LakshyaOjhaGitte> Stay safe!

15:55 < kartikdutt18Gitt> I hope everyone stays safe. Thanks.

16:57 < kartikdutt18Gitt> Hey @rcurtin, mlpack-bot doesn't seem to label issues and PR anymore. Reference tickets #2329 and #2330.

17:04 < rcurtin> >:( another day debugging javascript

17:04 < rcurtin> let me look into it and try to figure it out :)

17:13 < kartikdutt18Gitt> Thanks a lot.

17:31 < kartikdutt18Gitt> Hey @rcurtin, @zoq, for downloading dataset during runtime in [#3](https://github.com/mlpack/models/pull/3), I am using wget, so whenever a dataset is needed or pertained weights is set to true it will be downloaded (else it won't be). Do you suggest some other way of doing this or is this okay?

17:33 < rcurtin> CMake can do it too :) https://cmake.org/cmake/help/latest/command/file.html

17:33 < rcurtin> see `file(DOWNLOAD ...)`

17:33 < rcurtin> looks like that's available for all CMake 3 versions, which is nice

17:33 < rcurtin> that could provide a platform-independent way to do it, but I don't know if it fits into the workflow you were thinking

17:35 < kartikdutt18Gitt> I think you once mentioned to avoid using cmake so I went with wget. I had already implemented it with Cmake in mlpack/examples#60 . What do you suggest?

18:00 < kartikdutt18Gitt> I think the advantage to something like wget is that when CLI is also added, in one file we could do `--DATASET=mnist` and mnist would be downloaded if it doesn't exist. The same file could be used to download any dataset that we support.

18:07 < rcurtin> sorry, maybe I misspoke; my suggestion was to avoid CMake for the examples repository, but for the models repository it seems to make sense to me

18:08 < rcurtin> maybe it makes sense to download straight from C++ then, and not use wget? or do you mean that you are using a C++ interface for wwget?

18:09 < rcurtin> *wget

18:15 < rcurtin> ahhh, ok, that makes more sense to me

18:15 < rcurtin> I thought you meant that this would be done during configuration/build

18:15 < rcurtin> maybe libcurl would be more widely available and installed?

18:15 < kartikdutt18Gitt> And yeah we need Cmake for models. Sorry I wasn't clear, I meant you suggested that maybe we could find an alternative to downloading weights and file using Cmake rather than doing something like `-DDOWNLOAD_DATASETS=ON` or `-DDOWNLOAD_MNIST=ON`

18:16 < rcurtin> agreed, I think downloading them at runtime is a reasonable idea

18:16 < rcurtin> it makes more sense for when these bindings are distributed

18:16 < kartikdutt18Gitt> Ahh no, that wouldn't be a very good idea. The user would be downloading very large datasets that he won't use.

18:17 < kartikdutt18Gitt> Great. I will switch to libcurl

18:17 < rcurtin> sounds good---sorry for any confusion on my part :)

18:19 < kartikdutt18Gitt> No concussion on my part, I think I am up to date. Will push a commit tomorrow for lib curl. Thanks a lot.

18:20 < kartikdutt18Gitt> *confusion

18:37 < rcurtin> hopefully no concussion also :-D

19:15 < kartikdutt18Gitt> LOL, no concussion as well 😂

23:23 < KarthikeyanR[m]> Hi, regarding your gsoc idea for ml pack on resource constrained devices such as raspberry pi. incidentally I had been developing a project called klib for the same purpose. Can I just use for gsoc and use that to complete the project and somehow merge it later?

23:32 < KarthikeyanR[m]> I am sorry if i am too late for gsoc but I am still working forward to contribute to the community