07:09 < kartikdutt18Gitt> Hey @saksham189, Would it be okay to discuss some proposal related questions here ? (This is in reference to a comment you left on my Proposal).

07:13 < kartikdutt18Gitt> > `zoq on Freenode` Prince Gupta (Gitter): Agreed, I think we can concat the input?

07:13 < kartikdutt18Gitt> I think that makes sense. I will add a comment to #2264 to move forward with the same. Thanks a lot.

10:05 < GauravTirodkar[m> Hello Everyone! My name is Gaurav, and I'm in my Junior year of Computer Engineering. I have a huge interest in programming and have been working around with python c++ for a long time. I also have worked on a few deep learning projects. Any sort of guidance on the application process would be really appreciated!

10:23 < Param-29Gitter[m> Hey Gaurav, I hope these are helpful: https://www.mlpack.org/community.html , https://www.mlpack.org/gsoc.html

11:09 < GauravTirodkar[m> Is a contribution compulsory before submitting the application

11:10 < GauravTirodkar[m> because I have done a lot of deep learning work but haven't contributed yet

11:11 < himanshu_pathak[> kartikdutt18 (Gitter): Hey can you tell on which project you are trying to work

11:13 < favre49> GauravTirodkar[m: Contribution gives us an idea of your knowledge of C++ or ML, but it's not compulsory at all

11:24 < kartikdutt18Gitt> Hey Himanshu, I have applied for implementation of ANN Algorithms , mainly adding an object detection suite.

11:27 < himanshu_pathak[> kartikdutt18 (Gitter): Oh good I have applied for implementing deep learning modules adding some classification modules

11:29 < kartikdutt18Gitt> Great, Both of these projects should be fun to work on.

11:34 < himanshu_pathak[> Yeah it's good to work for but we can complete only one in summer

11:34 < kartikdutt18Gitt> Also I made my way to Pooling layers for the copy constructor PR, at least till there I think we can call the other constructor and only problem is with recurrent layer and work around would involve storing the constructor params similar to what you did in Convolution, what do you suggest?

11:36 < himanshu_pathak[> kartikdutt18 (Gitter): yes we can do That

11:39 < himanshu_pathak[> I am just thinking about we can introduce some memory leak while calling a constructor inside another constructor may be I am wrong what do you suggest?

11:46 < GauravTirodkar[m> <himanshu_pathak[ "kartikdutt18 (Gitter): Hey can y"> I am really intrigued with the Rrinforcement learning project.

11:46 < GauravTirodkar[m> > <undefined>

11:46 < GauravTirodkar[m> (edited) ... intrigued with the Rrinforcement learning ... => ... intrigued by the Reinforcement learning ...

11:47 < himanshu_pathak[> kartikdutt18 (Gitter): If that's the case So, I am wrong

11:48 < himanshu_pathak[> May be we can do this

11:48 < himanshu_pathak[> Gaurav Tirodkar: Good what are you trying to add

11:48 < kartikdutt18Gitt> himanshu_pathak: I am not sure though, I think we can wait for a reply from @rcrutin, @zoq

11:49 < himanshu_pathak[> kartikdutt18 (Gitter): yeah we should wait for their comments

11:54 < kartikdutt18Gitt> Gaurav Tirodkar, That's very interesting, We have an amazing tutorial for that. You can take a look at this [link](https://github.com/mlpack/mlpack/tree/master/doc/tutorials/reinforcement_learning).

12:02 < saksham189Gitter> @kartikdutt18 I have replied to you on the proposal. I think it would be best to discuss on the proposal itself so that we can keep track of the discussion.

12:02 < saksham189Gitter> (edited) ... the discussion. => ... the discussion although it is also fine to ask any questions here.

12:04 < kartikdutt18Gitt> That makes sense, Thanks a lot.

14:13 < LakshyaOjhaGitte> In the >alpha Dropout layer

14:13 < LakshyaOjhaGitte> I can see in the definition in the paper that in a layer we try to push the activation result mean to 0 and variance to 1, but I am not able to decipher the same from the implementation. can someone help me with the same?

14:13 < LakshyaOjhaGitte> (edited) ... the >alpha Dropout ... => ... the alpha Dropout ...

15:16 < LakshyaOjhaGitte> maybe the documentation can help

15:40 < GauravTirodkar[m> <kartikdutt18Gitt "Gaurav Tirodkar, That's very int"> Yes! Definitely.

15:42 < ShikharJaiswalGi> @rcurtin Some proposals do not have edit access provided to them, so essentially I don't think that the mentors can comment on them. I have rather added my suggestions on internal review, but I don't think the students can see them.

15:44 < himanshu_pathak[> Shikhar Jaiswal (Gitter): Hey can you a review on my proposal I have to change many things but some comments from your side will bw helpful

15:44 < himanshu_pathak[> * Shikhar Jaiswal (Gitter): Hey can you give a review on my proposal I have to change many things but some comments from your side will be helpful

15:46 < himanshu_pathak[> Implementing Essential deep learning module

15:48 < ShikharJaiswalGi> @himanshupathak21061998 Sure.

15:49 < himanshu_pathak[> Shikhar Jaiswal (Gitter): Thanks

15:51 < kartikdutt18Gitt> Hey @ShikharJ, Can I trouble you for a review as well, if possible I would love to get your opinion as well. Its on Application of ANN Algorithms.

16:02 < ShikharJaiswalGi> @kartikdutt18 The priority is on unreviewed proposals for now, so I'm not sure if I'll have a chance to go through all, I'll try my best.

16:02 < zoq> ShikharJaiswalGi: Student can't see those comments, but I think since the instructions are pretty clear how to setup the appliction, there is nothing really we can do.

16:04 < ShikharJaiswalGi> @zoq I see, that's unfortunate.

16:05 < abhisaphire[m]> Hey Shikhar Jaiswal (Gitter), if you are comfortable with binding generators. Can you please have a look at my proposal also. Thanks and sorry if I am causing any trouble

16:05 < kartikdutt18Gitt> @ShikharJ, I understand, Thanks a lot.

16:18 < chopper_inbound4> Hi Shikhar Jaiswal (Gitter) zoq : I think I was the culprit. I have changed the access rights. I am really sorry about that. If possible can any of you revisit the application with name "Mrityunjay Tripathi".

16:24 < rcurtin> favre49: you here? maybe easier to debug the ensmallen push issue over chat :)

16:25 < favre49> rcurtin: yup

16:25 < rcurtin> so you tried to push a commit and it said the same thing about two approving reviews required?

16:25 < favre49> Yeah, same error

16:25 < rcurtin> ok... the Contributors team has "write" access... let me check the protected status of the branch

16:26 < rcurtin> oh, I see... I have "Require pull request reviews before merging", which says:

16:26 < rcurtin> "When enabled, all commits must be made to a non-protected branch and submitted via a pull request with the required number of approving reviews and no changes requested before it can be merged into a branch that matches this rule."

16:27 < rcurtin> but I don't have "Include administrators" checked, so the requirements never applied to me :-O

16:27 < rcurtin> so, I guess, it's actually me that has to change my workflow :)

16:27 < rcurtin> would it be easy for you to open a PR with the changes? I'm realizing now what we could have done that would work was push to Joe's branch as a maintainer (which I don't usually do), then merged

16:28 < rcurtin> I'll have to go ahead and do this release using my special administrator privileges, then I'll have to disable those special administrator privileges and figure out how to adapt the ensmallen-release.sh script accordingly :)

16:29 < rcurtin> I guess now, instead of pushing to master, it will have to open a PR for release changes (which is fine), and then I have to figure out how tags can be pushed (maybe those aren't protected)

16:29 < favre49> Ahh makes sense

16:29 < rcurtin> alternately, if you want, if you can just email me the patch from the commit (maybe made with `git format-patch`) I can just apply it to master and go from there

16:30 < favre49> Hmm let me give that a shot, never done that before. I'll mail you in a couple

16:31 < rcurtin> ok, sounds good

16:36 < favre49> Alright, should have been sent?

16:41 < rcurtin> haven't received it yet

16:46 < favre49> Ah, seems like a gmail issue. I'll try mailing it manually instead of from the cli

16:46 < favre49> That works, right?

16:47 < rcurtin> sure, I don't see why it wouldn't work :)

16:47 < rcurtin> if we get desperate you can just post it on pastebin :-D

16:49 < favre49> Alright, this time it should be sent

16:49 < favre49> I find mail infuriating in general, I'm currently trying to get mutt working on my laptop

16:52 < rcurtin> favre49: thanks, the format wasn't quite right from what I got for `git am` but I could still see the changes so I just applied them :)

16:52 < rcurtin> next I'll release ensmallen, then start thinking about how to update that release script

16:53 < favre49> Sorry about that, I'll get the thing fixed. My list of things to fix before lockdown ends grows ever longer...

16:55 < rcurtin> no worries :)

16:56 < rcurtin> before we got locked down here I quickly went and bought some wood to build a desk

16:56 < rcurtin> only problem is, I wasn't able to go by a friend's wood shop to mill the wood, so I have to use hand planes

16:56 < rcurtin> :)

17:01 < jeffin143[m]> I love lockdown :)

17:05 < favre49> Sounds like a good workout

17:05 < rcurtin> yeah, my arms certainly get tired... :)

17:06 < rcurtin> it's a good routine... hand plane until the arms get tired, come back inside and type until the fingers get tired, rinse and repeat :)

17:06 < favre49> jeffin143[m]: The only thing that's changed for me is that I'm at home.. I still have online classes

17:13 < jeffin143[m]> favre49: omg

17:13 < jeffin143[m]> They are teaching you ?

17:13 < jeffin143[m]> I don't find Indian institutes teaching that good 😂

17:14 < favre49> Yup. BITS has no attendance policy though, so it's not often that I go for them

17:15 < favre49> Unless the subject is taught well. All the classes are recorded, so I can watch them at my own leisure

17:21 < travis-ci> mlpack/ensmallen#767 (2.12.0 - 12a2522 : Ryan Curtin): The build passed.

17:21 < travis-ci> Change view : https://github.com/mlpack/ensmallen/compare/a1c65c55d5bf^...12a2522df7cf

17:21 < travis-ci> Build details : https://travis-ci.org/mlpack/ensmallen/builds/668123806

17:24 < LakshyaOjhaGitte> Same goes here apparently dtu is also able to pull off online classes. Cant say anything how much this lockdown will extend

17:31 < LakshyaOjhaGitte> @rcurtin you here?

17:47 < rcurtin> LakshyaOhjaGitte: sort of, I was about to go get up and handle some other things

18:10 < LakshyaOjhaGitte> oh, no prob will ping you tomorrow going to bed rn

18:11 < LakshyaOjhaGitte> also can you review my draft proposal?

19:03 < rcurtin> I'll see if I can get to it, no promises

19:54 < jeffin143[m]> Y elem_type is not documented in armadillo docs

19:54 < jeffin143[m]> There are only 2 instances of it

21:02 < travis-ci> shrit/examples#20 (digit - c99fb8d : Omar Shrit): The build is still failing.

21:02 < travis-ci> Change view : https://github.com/shrit/examples/compare/ca9559de4eb2...c99fb8d976f8

21:02 < travis-ci> Build details : https://travis-ci.com/shrit/examples/builds/156187284

21:11 < travis-ci> shrit/examples#21 (master - 5eee8ab : Ryan Curtin): The build passed.

21:11 < travis-ci> Change view : https://github.com/shrit/examples/compare/6285a65983b0...5eee8abd328b

21:11 < travis-ci> Build details : https://travis-ci.com/shrit/examples/builds/156191387

21:43 < travis-ci> shrit/examples#22 (digit - 7fddd16 : Omar Shrit): The build was fixed.

21:43 < travis-ci> Change view : https://github.com/shrit/examples/compare/c99fb8d976f8...7fddd163faaa

21:43 < travis-ci> Build details : https://travis-ci.com/shrit/examples/builds/156228731

22:00 < himanshu_pathak[> Hey zoq: I am thinking adding test I will try to run 10 iteration of altering gibbs sampling will it be sufficient what do you suggest?

22:00 < himanshu_pathak[> To test DBN

22:00 < zoq> himanshu_pathak[: Sounds reasonable, yes.

22:01 < himanshu_pathak[> zoq: Ok By tommorow I will update my proposal thanks for suggestion

23:29 < rcurtin> jeffin143[m]: have you had a chance to look at #2019 yet? I think it's close to ready to merge---I want to see if we can keep the 3.3.0 release moving :)