ChanServ changed the topic of #mlpack to: "mlpack: a fast, flexible machine learning library :: We don't always respond instantly, but we will respond; please be patient :: Logs at http://www.mlpack.org/irc/
travis-ci has joined #mlpack
< travis-ci> mlpack/ensmallen#1058 (master - a37f49c : Ryan Birmingham): The build was broken.
travis-ci has left #mlpack []
Supr3me-code has joined #mlpack
Supr3me-code has quit [Remote host closed the connection]
ImQ009 has joined #mlpack
< say4n> zoq: I see. I'll go through the cpplint repo. :)
< yashwants19[m]> Hey [zoq](https://matrix.to/#/@freenode_zoq:matrix.org) [rcurtin](https://matrix.to/#/@freenode_rcurtin:matrix.org) should we remove Linux R and macos R build from azure pipeline as we are testing R-bindings for windows, linux and macos from github actions?
< zoq> yashwants19[m]: Yeah, don't see a reason to run the build/test twice, I think the only downside is that now a user has to look at two different systems.
< yashwants19[m]> Yeah, that's one case.
< zoq> yashwants19[m]: I guess at some point we could push the azure config to GitHub Actions.
< yashwants19[m]> Yeah, We could think about this.
< rcurtin> zoq: yashwants19[m]: doesn't github actions have a maximum number of build minutes per month that's way lower than azure?
< zoq> Ohh, yeah you are right, that's why we can't switch.
< rcurtin> yeah I am sad about that because it would be really nice to have all the configs in one place :(
< zoq> I guess you can put the config to azure, but since it currently works I think there is no real benefit to do that.
BlakJak888[m] has quit [Ping timeout: 256 seconds]
BlakJak888[m] has joined #mlpack
< rcurtin> shrit[m]: ok, outlook claims that they whitelisted my IP... it seems like emails now get through, but I don't know whether they end up in the spam folder or anything
< rcurtin> I subscribed to the mlpack-git list and at least the subscribe/confirm emails didn't go to the spam folder
< shrit[m]> Perfect
< shrit[m]> Maybe some test to outlook address.
< rcurtin> sent a test to your outlook address :)
BlakJak888[m] has quit [Ping timeout: 240 seconds]
< himanshu_pathak[> Hey kartikdutt18 (Gitter) was `convLayer.Parameters().n_elem` working on your pr I think I saw all build green on your pr https://github.com/mlpack/mlpack/pull/2637
< himanshu_pathak[> Or may be I am wrong :)
< himanshu_pathak[> * Hey kartikdutt18 (Gitter) was `convLayer.Parameters().n_elem` working on your pr I think I saw all previous build green on your pr https://github.com/mlpack/mlpack/pull/2637
< abernauer[m]> zoq: Got through a decent chunk of rewriting the example code yesterday,
< kartikdutt18[m]> Hey himanshu_pathak , The Parameters().n_elem didn't work. The previous green build was same as yours.
< himanshu_pathak[> Parameters is not accessible we are initializing layers using `LayerTypes` another way to use it `LSTM<> lstm(randomSize, randomSize);` but we will not be able to use WeightSizeVisitor like this
< jeffin143[m]> @zoq : mac os builds are failing
< jeffin143[m]> Due to ssl error
< jeffin143[m]> Can't connect to pypi
< rcurtin> jeffin143[m]: sounds unlucky, maybe it will be fixed in a few hours?
< jeffin143[m]> Oh I thought it was from our side
< jeffin143[m]> Got it , let's wait
< rcurtin> oh, hm, can I get a link to a failing build log?
< jeffin143[m]> Failed 24 times across 86 runs
< jeffin143[m]> Not sure if that is ok ?? Like once in 3 runs
< rcurtin> "Can't connect to HTTPS URL because the SSL module is not available."
< rcurtin> maybe that is on us?
< jeffin143[m]> Yes , seeing the error I thought it was on us
< jeffin143[m]> And hence want to ask zoq suggestions
< rcurtin> want to try opening a PR with something like the fix here?
< jeffin143[m]> I can take a look later today night may be
< rcurtin> seems simple, just `brew uninstall --force openssl` and `brew install openssl`, maybe that will fix it?
< rcurtin> thanks for pointing it out, I didn't even know anything was wrong :)
< jeffin143[m]> I will take care :)
< jeffin143[m]> Just waiting for shrit to finish with cereal
< jeffin143[m]> To start migration of serialisation test , otherwise it woul be really tough for him to resolve conflicts , haha
< rcurtin> yeah; I think we are getting close... just need to try disabling cotire and if that still doesn't work, abandon VS15
< jeffin143[m]> Oh if someone uses vs15 then ?
< jeffin143[m]> Wouldn't it create some issues ??
< rcurtin> I think that VS15 will still work... if you have a massive amount of RAM on your machine...
< rcurtin> but the azure build nodes don't have enough memory to actually run the mlpack build
< rcurtin> I suspect that once we transition completely away from boost, that we could re-enable VS15... but that might be a little while
< jeffin143[m]> Massive ?? In theses ?? Any base limit
< jeffin143[m]> 8 gb ram ??
< rcurtin> no idea what the azure nodes currently have; maybe 8GB or 16GB is necessary?
< rcurtin> in any case, VS16 seems to handle it fine
< rcurtin> so if a user complains we can say for now that they can use VS16
< jeffin143[m]> Oh understood
< jeffin143[m]> Got it :) thanks for speeding me up to the real issue
< rcurtin> yeah, I am really hoping we can find some workaround, but it's possible we might not
< rcurtin> visual studio is tough to work with...
< jeffin143[m]> I think if we remove boost serialisation and boost testing that might help
< jeffin143[m]> Let's see
< rcurtin> agreed, when we can finally remove the unit test framework that should help too
< himanshu_pathak[> <rcurtin "agreed, when we can finally remo"> 🥳
< himanshu_pathak[> Hey kartikdutt18 Using `ParametersVistior` will work for that issue
< kartikdutt18[m]> Ohh, Awesome. Will make the fix. Thanks.
< himanshu_pathak[> <kartikdutt18[m] "Ohh, Awesome. Will make the fix."> 👍️
< shrit[m]> I have already disabled cotire,
< shrit[m]> I think the build for this one should be finished. I made a push 4 hours a go
< rcurtin> the build failed; I don't think you disabled cotire correctly :(
< jeffin143[m]> Yes
< jeffin143[m]> Some header issues
< jeffin143[m]> Log was shown as undeclared identifier
< rcurtin> I believe that cotire makes some changes so that prereqs.hpp is always included in all files, but now that cotire is disabled, that does not work correctly anymore
< rcurtin> so, it may be that some headers need prereqs.hpp added to them or something
< rcurtin> or it may be something else
< shrit[m]> Ah I see.
< shrit[m]> because I only need to comment cotire in order to disable it
< shrit[m]> unless if there is any other thing that I do not know
< rcurtin> you might want to look at the git log of when cotire was originally enabled, to see if there were other changes being made there?
< rcurtin> or read the cotire documentation to see maybe what it might be doing
< rcurtin> set_target_properties(mlpack PROPERTIES COTIRE_CXX_PREFIX_HEADER_INIT "prereqs.hpp")
< rcurtin> that's the relevant line, I think
< rcurtin> you should be able to run this quickly locally to see what is going on
< shrit[m]> of courser
< shrit[m]> I am running it locally to see what is happenning.
< AakashkaushikGi4> Hey all, this is just a general question about what will be the major goal for mlpack in the next year or so, I went through this https://github.com/mlpack/mlpack/issues/2524#issuecomment-662692857 but to be honest as a beginner it's not really obvious for me to know where to concentrate my efforts where they have the greatest impact.
< AakashkaushikGi4> The only precise guide i think i can follow helping with these issues [here](https://github.com/mlpack/mlpack/milestone/30).
< AakashkaushikGi4> (edited) ... follow helping ... => ... follow is by helping ...
< rcurtin> AakashKaushikGi4: I'm in the process of writing up something more comprehensive related to #2524, but it's not ready yet (and I'm not sure it will fully help to answer your question)
< rcurtin> I think, maybe the best thing to do, is take a look at the #2524 discussion, see what people seem to think is important, and try to understand the underlying things that would need to be done
< rcurtin> I'm sorry if that's not particularly prescriptive...
< rcurtin> in some sense, it's up to you to figure out what you think is most interesting to work on :-D
< rcurtin> certainly basically anything you might do would be helpful to the library overall!
< AakashkaushikGi4> No, I definitely understand the complete situation and i almost went through the complete discussion except the part where people were talking about GPUs specifically . I think i will give it a read again and try to figure out something.
< abernauer[m]> Aakash kaushik (Gitter): I would say spend time on an issue you find interesting and ideally is beginner friendly. If you're having trouble on the issue ask for help or work on your knowledge that is relevant to the issue.
< AakashkaushikGi4> @abernauer thanks for the guidance, I would keep that in mind moving forward. : )
ImQ009 has quit [Quit: Leaving]
sreenik[m] has quit [Ping timeout: 244 seconds]
AbishaiEbenezer4 has quit [Ping timeout: 244 seconds]
yc7871[m] has quit [Ping timeout: 244 seconds]
Saksham[m] has quit [Ping timeout: 244 seconds]
vigsterkr[m]1 has quit [Ping timeout: 244 seconds]
yuvraj_2701[m] has quit [Ping timeout: 244 seconds]
jeffinsamGitter[ has quit [Ping timeout: 244 seconds]
jonpsy[m] has quit [Ping timeout: 244 seconds]
AbishaiEbenezer7 has joined #mlpack
sreenik[m] has joined #mlpack
Saksham[m] has joined #mlpack
yc7871[m] has joined #mlpack
yuvraj_2701[m] has joined #mlpack
jeffinsamGitter[ has joined #mlpack
jonpsy[m] has joined #mlpack
vigsterkr[m]1 has joined #mlpack
nishantkr18[m] has joined #mlpack
< abernauer[m]> yashwants19: Thought there might be a bug with the preprocess describe function, but the verbose argument is not optional.
< zoq> jeffin143[m]: Fixed the macOS SSL issue in #2660 unfortunately none of the workaround in the azure post worked, so I ended up upgrading the macOS image.