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/
< AakashkaushikGit>
Everyone seems to be packed in October.
< CharanpreetSingh>
But pip install mlpack gives : ERROR: Could not find a version that satisfies the requirement mlpack (from versions: none)
< CharanpreetSingh>
ERROR: No matching distribution found for mlpack
< AakashkaushikGit>
try with `pip3 install mlpack`
< AakashkaushikGit>
also you can go here: https://pypi.org/project/mlpack/ and check if a windows version is available for any mlpack version, and if not you can just build the python bindings
< AakashkaushikGit>
you can check that in the download files section
< AakashkaushikGit>
also I don't think we have a mlpack wheel for windows through pip.
ImQ009 has quit [Read error: Connection reset by peer]
ImQ009 has joined #mlpack
kyrre has joined #mlpack
tpastorello has joined #mlpack
< tpastorello>
Hello guys. I have a problem compiling, I have the error: ... test1.cpp: In function ‘void mlpackMain()’:
< tpastorello>
test1.cpp:36:18: error: ‘GetParam’ is not a member of ‘CLI’; did you mean ‘mlpack::bindings::cli::GetParam’?
< rcurtin>
tpastorello: are you using git master? try replacing CLI::GetParam<> with IO::GetParam<>
< rcurtin>
recently we did a refactoring there and the name of the class changed
< rcurtin>
sorry for any inconvenience :)
< tpastorello>
hello rcurtin ... i'm using the example you sent me in the email! By the way thank you very much!
< tpastorello>
i will try IO:: now :-)
< rcurtin>
ah, sorry about that, I guess it must have been a bit out of date :)
< rcurtin>
glad that the resources are helpful!
< tpastorello>
now worked with IO:
< tpastorello>
is there an 'automatic' way to check the summary of the decision tree, accuracy, confusion table, etc.?
< rcurtin>
there's a function called ConfusionMatrix in src/mlpack/core/data/confusion_matrix.hpp that I think could be helpful
< rcurtin>
but not an all-on-one solution
< tpastorello>
thank you very much, i will try :-)
petris has joined #mlpack
petris has quit [Client Quit]
petris has joined #mlpack
< RishabhGarg108Gi>
@tpastorello , you can use confusion matrix(as suggested by @rcurtin) or F1-score if you are doing classification or you can use some accuracy measure like rmse/mse/R-score etc for regression task.
< tpastorello>
@RishabhGarg108Gi Thanks for the suggestion, I'm trying here. I will look for the documentation about the F1-score in mlpack
theloudspeaker has quit [Quit: Idle for 30+ days]
< RishabhGarg108Gi>
:)
kyrre has quit [Quit: Connection closed for inactivity]
< rcurtin>
I think the two PRs I suggested for mlpack 3.4.2 are a bit too far away from merge... any objections if I just release 3.4.2 today?
< zoq>
rcurtin: Fine with me.
< rcurtin>
ok, let me get the release process started :)
< rcurtin>
I think I'll also try to set up the mlpack.org website as a Jenkins job
< rcurtin>
that will be a step towards the same release process as ensmallen
< zoq>
cool
< CharanpreetSingh>
@Aakash-kaushik , even I think there is no mlpack wheel for windows through pip, also I tried pip3 install mlpack but it renders me the same error, can you elaborate on how I can "build the python bindings " for mlpack.
< AakashkaushikGit>
I think you should be pretty good with these two things, but let me know if you have any errors.
< AakashkaushikGit>
> `birm` Aakash kaushik (Gitter): yes. So, RishabhGarg108 (Gitter) pass the option to enable python binding with with -D BUILD_PYTHON_BINDINGS=ON on the cmake step. Then, see if the cmake output seems satisfied with the dependencies.
< AakashkaushikGit>
Also @CharanpreetSingh056 read the messages from here, I think this will be able to help you too.
< CharanpreetSingh>
Hey @Aakash-kaushik , I read the chats but still cannot figure it out, the main reason being I built it trrough vcpkg which required just two commands ( 1. PS> .\vcpkg install mlpack:x64-windows 2 . PS> .\vcpkg install mlpack[tools]:x64-windows ) , and "install integrate " to link it to my VS2019 , I had no step involving cmake.
< CharanpreetSingh>
But I have a vcpkg.cmake file here (C:/dev/vcpkg/scripts/buildsystems/vcpkg.cmake), are any changes to be made here ?
< rcurtin>
CharanpreetSingh: I've been trying to build the windows wheel for months to upload it to pip... it's a nightmare
< rcurtin>
I might not manage to release 3.4.2 today---I realized to get closer to automation, I need Jenkins to be building the website, but that takes a long time :)
< rcurtin>
so it might be tomorrow
< rcurtin>
or Monday if things go poorly :)
< rcurtin>
(of course I don't think it's a huge difference either way)
< zoq>
Maybe at some point we just have to run a script :)
< rcurtin>
yeah, I hope so! and I think it can be soon
< rcurtin>
once Jenkins can build the mlpack.org website if I just push a commit to the mlpack.org repo, I can have mlpack-bot do the exact same thing it already does for ensmallen
< rcurtin>
however, that won't update, e.g., downstream Julia/pip/conda/debian packages, so that step will have to be done manually still
< CharanpreetSingh>
Thanks @rcurtin ( and all others ) for your contributions to mlpack to make it easily accessible to the users , eagerly waiting for Monday ( or tomorrow ) !
< CharanpreetSingh>
Also @rcurtin , will the 3.4.2 version be a windows wheel or will it be a macOS wheel ?
< rcurtin>
CharanpreetSingh: I have scripts working that will build a macOS and Linux wheel, but like I said the Windows scripts are not working at the moment
< rcurtin>
I hope to be able to fix it soon... but like I said I have been trying (on and off) for months; it's very hard to figure out what is going wrong from the build logs