<Aakash-kaushikAa>
the license year seems to be expired.
<Aakash-kaushikAa>
2007 - 2020
VarunBG[m] has joined #mlpack
mlpackrocks has joined #mlpack
mlpackrocks has quit [Client Quit]
<shashankshet[m]>
Could anyone here point me to some resources on how to setup my development environment?
<shashankshet[m]>
I have built mlpack from source, are there any other steps to be followed?
<rcurtin[m]>
Aakash-kaushik (Aakash kaushik): thanks for pointing that out---I just pushed some commits that should hopefully fix that 👍️ (I guess we have to wait for jenkins to rebuild it)
<rcurtin[m]>
ok! looks better now 😄 let me know if you see any other issues (or you can also open a PR in the mlpack.org repository too)
<Aakash-kaushikAa>
Hey @ryan:ratml.org thanks for fixing that and sure will do that if I notice something else.
<zoq[m]1>
<shashankshet[m]> "Could anyone here point me to..." <- A lot of people like VS Code, so I think if you search for VS Code + C++ you will find some good tutorials.
<shashankshet[m]>
No, not the IDE per se. I am comfortable using emacs right now. So after building from source, there do not seem to be any instructions regarding how the setup works with git,
<shashankshet[m]>
So, do I maintain a copy of the code in a repo, in parallel with what I have just built?
<zoq[m]1>
Not sure I get the question, there is nothing special about mlpack in that regard, you you like to contribute you fork the repo on github, clone your fork, checkout a new branch and build that branch using (cmake, make). You can skip the `make install` step since for testing you can just do something like `bin/mlpack_test "[pca]"`.
<zoq[m]1>
But in essence you should build your own fork, and keep that fork up to date.
<shashankshet[m]>
Okay,
<shashankshet[m]>
So I do not need to download a tar file or anything fancy,
<shashankshet[m]>
* file or do anything fancy,
<zoq[m]1>
You can go to https://github.com/mlpack/mlpack click the fork button in the upper right, this will create a fork of mlpack and you can then clone that with git.
<rcurtin[m]>
Hey Abhi , thanks for pointing that out. for the version that appends the old PYTHONPATH with `:`, can you print the resulting `$PYTHONPATH`?
<rcurtin[m]>
or, you were saying that `build/include/` works, but `build/src/` does not?
<rcurtin[m]>
okay---so it's `include` vs. `src` that is the issue, not the `:` 👍️
<rcurtin[m]>
but what is in `build/src/mlpack/bindings/python/`, and what is in `build/include/mlpack/bindings/python/`?
<rcurtin[m]>
it should be the `src` directory that has all the compiled bindings
<rcurtin[m]>
also it is worth checking what happens when you leave `PYTHONPATH` unset (e.g. `export PYTHONPATH=`)
<shashankshet[m]>
zoq: Thanks a bunch
<shashankshet[m]>
* a bunch!
<rcurtin[m]>
Abhi: right, but does `build/src/mlpack/bindings/python/` have all the compiled `.so` files? (maybe in a subdirectory)
<rcurtin[m]>
Abhi: when you built, did you run `make python`? (what happens if you run that now?)
<rcurtin[m]>
oh, interesting, did you run `make python` from the root `build/` directory?
<rcurtin[m]>
👍️ thanks for all the diagnostic information :)
<rcurtin[m]>
it seems like CMake has not configured your system to build the python bindings---so you may try forcing them to be enabled with `cmake -DBUILD_PYTHON_BINDINGS=ON ../`
<rcurtin[m]>
there doesn't need to be, but you can add one---it's optional, CMake parses it just fine either way 👍️
<kuries[m]>
oh my bad, I forgot to include an underscore for one of them
<rcurtin[m]>
Binesh Munukurthi: 👍️ I think CMake will tell you that a variable was ignored if it was never used. close inspection of the configuration output can be helpful to determine whether or not, e.g., Python bindings will be compiled
<rcurtin[m]>
if you just reconfigured to set `BUILD_PYTHON_BINDINGS` to `ON`, then when you type `make` (or `make python`) CMake will not rebuild anything unnecessarily
<rcurtin[m]>
but be warned... compiling the Python bindings can take a long time, even if the base library's compilation is not being done 😄
<rcurtin[m]>
yep, use `-DPYTHON_EXECUTABLE=$(which python3)`
<rcurtin[m]>
(or whatever the path to the python executable is)
<rcurtin[m]>
zoq: I'm trying to set up Azure Pipelines for the examples repo... but I think that it's only possible to do this once the configuration is merged (judging by the interface I am looking at)
<rcurtin[m]>
yeah, so it seems like we should just merge https://github.com/mlpack/examples/pull/187 and then I can debug it from there? I think once it's merged I can tell Azure what file to look in, and then hopefully it will work 🤞
<zoq[m]1>
Right, we just need an initial build pipeline config in the repo.
<rcurtin[m]>
👍️ do you mind reviewing #187 then, and we can get it merged as a starting point?
<rcurtin[m]>
oh I see you already did 😄
<zoq[m]1>
already approved from my side
<rcurtin[m]>
any other maintainer willing to review approve it so we don't have to wait on mlpack-bot? 😄
<zoq[m]1>
I guess in theory you can just approve it yourself as well?
<zoq[m]1>
I don't see a section in the contributing document that prevents us from doing so
<rcurtin[m]>
Github doesn't allow me to, since I opened the PR :)
<zoq[m]1>
Ohh, I guess it wouldn't make sense.
<NabanitaDash[m]>
I am having issues with writing tests for copy & move constructor for adaptive mean and adaptive max pool. https://github.com/mlpack/mlpack/pull/3128 Can someone suggest some sources that I can get help from? I tried pytorch docs but not helpful.
<zoq[m]1>
<NabanitaDash[m]> "I am having issues with writing..." <- I'll take a closer look into the PR later today.
KunalA18 has joined #mlpack
KunalA18 has quit [Client Quit]
KunalA18KunalAga has joined #mlpack
<KunalA18KunalAga>
Good day, everyone! I'm Kunal Agarwal, a second-year undergraduate student. My domains of interest are Machine Learning/Deep Learning and OpenCV. I'd like to be a part of mlpack by making some valuable contributions. Any advice on how to get started and contribute would be extremely beneficial. Thanks