00:46 < jeffin143> lozhnikov : I need to maintain the count of token too, so in tf-idf policy , I have to declare , something as unordered_map<token_type,size_t> , but how to should I determine the token type in a policy?

00:49 < jeffin143> Should I take it as template parameter of class..???

00:50 < jeffin143> And then using tfidfencoding = stringencoding<tfidfencodingpolicy<tokenthpe>,stringdictionaryencoding<tokentype>>;

01:25 < jeffin143> lozhnikov : I just wanted to identify the token uniquely, so instead of storing token,I can store their value since it is same as that * , so I don't need token type

05:28 < lozhnikov> jeffin143: You don't need the token itself. It's enough to pass the label to the policy.

07:14 < Toshal> ShikharJ: Sorry for the slow response. I have completed my exact-objective PR. and I have worked on the visitors. I will push changes related to it soon. I have also made the changes you told me to do on the serialization PR.

07:14 < Toshal> Please review the serialization PR in your free time.

07:36 < ankit45> hello everyone, I am Ankit, a third-yearite at BITS Pilani, India. It is a pleasure to join this community. I would like to start contributing to mlpack. Could someone please point out some documentations that I could read so as to get a general Idea of how I can start contributing?

07:41 < jenkins-mlpack2> Project docker mlpack nightly build build #408: STILL UNSTABLE in 3 hr 27 min: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/408/

11:14 < xiaohong> ankit45: You can read the document, https://www.mlpack.org/doc/mlpack-3.1.1/doxygen/anntutorial.html.

11:14 < xiaohong> https://github.com/mlpack/mlpack/blob/master/CONTRIBUTING.md

13:33 < Toshal> ShikharJ: I am here

13:36 < ShikharJ> Toshal: Great, could you explain what you mean by exact-objective PR?

13:36 < ShikharJ> sakshamB: Are you there?

13:41 < Toshal> ShikharJ: it's this https://github.com/mlpack/ensmallen/pull/109

13:41 < rcurtin> Toshal: I think that's ready to go, just haven't had a chance to do the final review. I was hoping to find some time today

13:42 < rcurtin> thanks again for your hard work on that one by the way, it is really nice support :)

13:42 < Toshal> rcurtin: Thanks and no worries about the delay.

13:43 < Toshal> ShikharJ: I will make a PR for LSGANs soon. Sorry for the delay about it.

13:43 < ShikharJ> Toshal: Okay great, could you tell me which of your PRs are still under progress?

13:44 < ShikharJ> Maybe I can review the ones that are complete?

13:44 < Toshal> ShikharJ: Ah okay. Just a sec.

13:45 < Toshal> https://github.com/mlpack/mlpack/pull/1956

13:45 < Toshal> rcurtin: Can you just let me know your thoughts on https://github.com/mlpack/mlpack/pull/1924.

13:46 < rcurtin> hang on, let me pull it up...

13:46 < rcurtin> still failing the memory checks, I'm not sure why. let me re-run Jenkins on it and we can see what the error messages are

13:47 < rcurtin> the most recent time I looked at it, the memory checks job had just timed out, so I re-ran it, but didn't manage to go back and check the result

13:47 < Toshal> Okay. No worries.

13:48 < Toshal> ShikharJ: I am mentioning below PRs which you can review.

13:49 < Toshal> #1770, #1815, #1888 That's it.

13:49 < Toshal> I will complete work on #1956 and #1920 soon.

13:50 < Toshal> May be today itself.

13:50 < Toshal> I will ping you about them to you.

13:51 < ShikharJ> Toshal: Okay good. Please let me know as soon as they're complete.

13:51 < Toshal> #1939 and #1950 are also under progress.

13:51 < Toshal> They will need some time.

13:53 < Toshal> I guess #1933 just need discussion that's it.

13:53 < Toshal> Once you review it.

13:53 < Toshal> Not sure whether it's complete.

13:54 < ShikharJ> Toshal: Okay great, since you and I seem to have some catching up to do, we'll let this week slide on prior work. See you on Friday then. Have a good week.

13:55 < Toshal> ShikharJ: Okay Thanks.

13:56 < Toshal> Have nice day.

16:16 < jeffin14337> lozhnikov : I made a new branch from your branch, and add BOW and both TFIDF encoding algorthim

16:17 < jeffin14337> The only issue is TFidf Encoding has double values where, the Encode() function takes vector<vector<size_t>output as input

16:18 < jeffin14337> I will also handle the comments in string-cleaning PR and make a commit today

16:18 < zoq> xiaohong: Hello, did you see my comment on the PPO PR?

16:18 < lozhnikov> jeffin14337: Thanks, can you open a PR?

16:19 < jeffin14337> Also I will write the blog tomorrow :)

16:19 < lozhnikov> jeffin14337: Probably the type of the elements should be a template parameter.

16:19 < lozhnikov> jeffin14337: Ok

16:21 < sakshamB> ShikharJ: sorry for missing the meet. I hope to discuss on Friday or anytime before that as you like. My work on CGAN is almost complete so, you can review that.

17:16 < jeffin14337> lozhnikov : I have left some comments, PTAL when you are free

17:22 < jeffin14337> ok so i have understood the problem

17:22 < jeffin14337> The static data members are the problem

17:23 < jeffin14337> if I am calling the A::static function for a things

17:24 < jeffin14337> and again calling it, The intial state is not gone, How to flush out the old values ?

17:24 < jeffin14337> from a static data member?

17:25 < jeffin14337> lozhnikov : https://pastebin.com/yi6gdu3e

17:27 < lozhnikov> jeffin14337: What's that?

17:29 < lozhnikov> jeffin14337: Look at the TfIdf policy. It shouldn't contain static variables.

17:30 < jeffin14337> yes it shouldn't contain static variable

17:30 < jeffin14337> I just tried following the same function defination

17:30 < jeffin14337> I will just remove the static members, Thanks :)

18:47 < jeffin143> O

18:48 < jeffin143> lozhnikov : bow and tf-idf is ready , I have resolved the issue and added template parameter for vector<vector<>>

18:49 < lozhnikov> jeffin143: Okay, I'll look through the code today.

21:33 < abernauer> rcurtin: Any tips on dealing with SFINAE?

21:34 < rcurtin> abernauer: sure, I can give tips, but I need more context :)

21:34 < rcurtin> I wouldn't have imagined you would need to deal with SFINAE at all for just getting the handwritten R PCA binding working though

21:38 < abernauer> Well I was dealing with this in May too I think, but hadn't realized it. Caused me problems when working on CLI wrappers early. Also it's coming up when I try to pass a reference to an arma matrix as a function parameter.

22:42 < rcurtin> abernauer: if you're wrapping all of the CLI functionality as C functions, then you shouldn't be passing references around I don't think