03:40 < Victor_> hello there, I am a CS sophomore at the University of Rochester. I am interested in implementing a few Reinforcement Learning Algorithms for mlpack through GSoC. Specifically, I was interested in Proximal Policy Optimization and Persistent Advantage Learning DQN. I was wondering what kind of background are you looking for and what other metrics are you going to use to evaluate a proposal. Cheers, looking forward to hearing from you

05:12 < rcurtin> hi Victor_, take a look at http://www.mlpack.org/gsoc.html and hopefully that should help

05:12 < rcurtin> there is also an application guide linked to from the Ideas page, which perhaps might also help clarify :)

05:30 < sonu628> when i do sudo make install in the middle we get the following

05:30 < sonu628> TEST FAILED: usr/local/lib/python3.7/site-packages/ does NOT support .pth files error: bad install directory or PYTHONPATH

05:30 < sonu628> You are attempting to install a package to a directory that is not on PYTHONPATH and which Python does not read ".pth" files from. The installation directory you specified (via --install-dir, --prefix, or the distutils default setting) was: usr/local/lib/python3.7/site-packages/ and your PYTHONPATH environment variable currently contains: '/usr/local/lib/python3.7/site-packages/'

05:31 < sonu628> when i do $PTHONPATH in terminal i getbash: /usr/local/lib/python3.7/site-packages/: Is a directory"

05:31 < sonu628> if its a valid directory why is installation getting directed to usr/locao/...

05:31 < sonu628> any suggestions ?

07:47 < Viserion> I am not able to locate cf_model.hpp , cf_model_impl.hpp . It was suggested to look into them before solving the issue, I could see only cf_main.cpp and cf_impl.hpp

08:16 < sreenik> Viserion: I just checked the src/mlpack/methods/cf directory and there appears to be both cf_model.hpp and cf_model_impl.hpp. If you are searching the installation folder in usr/local/include/mlpack, you won't find them there. Just check your cloned repo directory

08:22 < jenkins-mlpack2> Project docker mlpack nightly build build #264: FAILURE in 1 day 4 hr: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/264/

08:23 < jenkins-mlpack2> Project docker mlpack nightly build build #265: STILL FAILING in 1 min 16 sec: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/265/

11:18 < jeffinsam> zoq : i guess there is some issue with jenkins, the checks for pull request have been stopped.i mean not running

12:10 < zoq> jeffinsam: Should be fixed, there are some jobs in the backlog so this might take some time.

12:45 < mulx10> zoq: I have a question regarding the issue (https://github.com/mlpack/mlpack/issues/1782). In the file lstm.hpp variables OutputDataType delta, OutputDataType grad, OutputDataType outputParameter; are not used anywhere. However they have getters and setters. So how should I get the `cell` parameter: (1) by writing getter and setters or (2) Function overloading (Forward(input, output, cell))

12:45 < mulx10> zoq: Thanks

12:54 < zoq> mulx10: They are used to hold e.g the error, the FFN class is one example, I guess to set the cell you could overload the Forward function.

14:17 < Nitish> can you explain ESO?

14:26 < mulx10> jeffinsam, zoq: I guess jenkins has stopped. It got stuck in test case.

14:27 < mulx10> Can someone please kill the process.

14:29 < mulx10> sorry for the trouble

14:46 < zoq> mulx10: Do you mean the memory test?

14:58 < mulx10> zoq: Yes

14:59 < zoq> mulx10: Okay, I've stoped the job.

15:02 < mulx10_> zoq: I guess the tests are fragile. I found the reason for my failing memory builds. ctest -j2 runs two parallel test. Also, a lot of tests uses the same file name, which in tur creates a race condition

15:03 < mulx10_> zoq: Also when I changed all the file names corresponding to the test case. It somehow got stuck.

15:03 < zoq> mulx10_: We could add the specific suite to https://github.com/mlpack/mlpack/blob/master/src/mlpack/tests/CMakeLists.txt#L181

15:04 < zoq> mulx10_: That might as well fix the issue with the same name.

15:08 < mulx10_> zoq: Thanks.

15:59 < jeffinsam> zoq : implemented one hot encoding function as utility function, any comments... would it be useful or not as preproccessing of data.?

16:28 < jeffinsam_> also anyone, any idea what is Static Code Analysis Checks, i couldn't get why pr is failling that test

16:29 < zoq> jeffinsam_: Which one is failing? Right now it runs a combination of cppcheck and pvs-studio.

16:30 < zoq> jeffinsam_: About the one hot encoding, definitely interesting, it's part of the String Processing Utilities idea (https://github.com/mlpack/mlpack/wiki/SummerOfCodeIdeas#string-processing-utilities)

16:31 < jeffinsam_> zoq : pr#1780

16:33 < zoq> jeffinsam_: I see, the node that runs the job doesn't use the right toolchain.

16:34 < zoq> jeffinsam_: Let's restart the test, but it might take some time, since the node is currently testing the docker suite.

16:35 < jeffinsam_> thanks zoq : i have make a pr regarding one hot encoding pr#1784 , could you review it whenever you get time, being a competitive programmer i am facing tough time writting code regarding industries standards

16:35 < jeffinsam_> zoq : no issues with the time , Thanks for the clarification

16:36 < zoq> jeffinsam_: I'll take a look once I have the chance.

16:38 < jeffinsam_> zoq : thanks

16:59 < Shubhangi100> Can anybody help me out through this " Thread 1 "python" received signal SIGSEGV, Segmentation fault, arma::Mat<unsigned long>::steal_mem (x=..., this=0x153f3d0) at /usr/include/armadillo_bits/Mat_meat.hpp: at line 1094 if(t_vec_state == x_vec_state)"

17:01 < ShikharJ> Shubhangi100: Can you mention the context of this error? What exactly are you trying to do?

17:04 < Shubhangi100> I am trying to debug segmentation fault with the help of python gdb in issue 'https://github.com/mlpack/mlpack/issues/1772'

17:10 < ShikharJ> Hmm, have you set the breakpoints at the required place yet? It seems that the program is seg-faulting before any of the set break-points.

17:25 < Shubhangi100> No, I haven't yet set any breakpoint.

18:21 < ShikharJ> Shubhangi100: You should set break-point before running the program in gdb, and after the break-point run it line by line to get to the point of error on the outer code.

18:41 < rcurtin> Shubhangi100: also use 'backtrace' and then move around in the stack frames (like 'frame 3' or so once you've seen the backtrace) and inspect the variables (i.e. 'inspect t_vec_state') in order to see what's going on

18:41 < rcurtin> I'd suggest trying to find the last mlpack-related stack frame and work from there

18:41 < rcurtin> hope this helps; there's likely to be a lot of reasoning necessary to understand what's happening and why it's a problem

20:35 < Viserion> what is the difference between dense matrix and sparse matrix of user item table in cf_impl.hpp in Collaborative filtering

21:39 < last_comer> hello

21:40 < last_comer> i am new to open source but interested in contributing

21:40 < last_comer> need help

21:41 < zoq> last_comer: Hello, http://mlpack.org/involved.html should be helpful here.

21:41 < last_comer> what is this?

21:41 < zoq> last_comer: http://mlpack.org/gsoc.html could be helpful as well.

21:42 < zoq> A page with a bunch of helpful infos about how to get involved and contribute.

21:44 < last_comer> what i supposed to know in order to contribute?

21:47 < zoq> That highly depends on what you like to do, but if you are willing to learn no stuff you are basically good to go.