verne.freenode.net changed the topic of #mlpack to: http://www.mlpack.org/ -- We don't respond instantly... but we will respond. Give it a few minutes. Or hours. -- Channel logs: http://www.mlpack.org/irc/
chennakeshava has joined #mlpack
mrcode has joined #mlpack
mrcode has quit [Client Quit]
chennakeshava has quit [Quit: Page closed]
wenhao has joined #mlpack
govg has quit [Ping timeout: 260 seconds]
wenhao has quit [Ping timeout: 260 seconds]
govg has joined #mlpack
eddyrene has quit [Quit: Page closed]
adi_ has joined #mlpack
< adi_>
hello everyone
< adi_>
my name is Aditya Sinha from 2nd year computer science engineering
< adi_>
I want to participate in GSoC this year
< adi_>
i m interested in String processing utilities project...
< rcurtin>
hi adi_, nice to meet you
< rcurtin>
I'm sorry for the quick response but I am about to go to bed...
sumedhghaisas has quit [Read error: Connection reset by peer]
sumedhghaisas has joined #mlpack
sumedhghaisas2 has joined #mlpack
sumedhghaisas2 has quit [Read error: Connection reset by peer]
sumedhghaisas3 has joined #mlpack
sumedhghaisas has quit [Ping timeout: 256 seconds]
aman____ has quit [Quit: Connection closed for inactivity]
sumedhghaisas has joined #mlpack
sumedhghaisas3 has quit [Ping timeout: 256 seconds]
sumedhghaisas2 has joined #mlpack
sumedhghaisas has quit [Ping timeout: 265 seconds]
moksh has joined #mlpack
< moksh>
@zoq, really sorry for the late reply. I just saw that travis-ci build failed. I'll try fixing that as soon as possible
adi_ has quit [Quit: Page closed]
< zoq>
moksh: No worries, we don't have to rush.
adi_ has joined #mlpack
< moksh>
@zoq, I tried running the tests locally to check the stability, but they run without any errors.
< zoq>
moksh: And you used another random seed for each run?
< adi_>
hello rcurtin..
< adi_>
i hv downloaded mlpack and started tutorial..but there seems some problem!
< rcurtin>
adi_: I can try and help, but you are going to need to give more details than that
< adi_>
yes yes sir...mlpack_logistic_regression -t dataset.csv -v the above code is resulting to segmentation fault..
< rcurtin>
okay, can you tell me what you have tried to resolve the issue?
< adi_>
actually i tried to search the dataset.csv file but i was not able to find it...so i instead put iris.csv in terminal ...
< adi_>
that also gave same result..
< moksh>
@zoq I used math::RandomSeed(std::time(NULL)); as mentioned in the ticket you referenced
< zoq>
moksh: Okay, can you build with -DDEBUG=ON and run the test again?
< adi_>
@rcurtin can i skip tutorial part and instead start making simple mlpack programs as meintioned in involved.html page?
< zoq>
adi_: Going throught the tutorials is an improtatnt step, so we should see if we can fix the issue.
< zoq>
*through
< zoq>
adi_: Is libmlpack in your search path?
< adi_>
but @zoq datasets are not being loaded ....its giving segmentation fault core dumped!
< moksh>
@zoq, Is this correct: bin/mlpack_test -DDEBUG=ON -t AdamTest/AdamirrorLogisticRegressionTest ? Because it gives an error
< zoq>
yeah, I read that part, still you ahve to make sure libmlpack.so is in your library search path
< zoq>
also what is the output of: 'mlpack_logistic_regression -h'
< zoq>
moksh: You have to rebuild mlpack: 'cmake -DDDEBUG=ON ..'
< adi_>
same segmentation fault
< rcurtin>
sorry, I stepped out
< rcurtin>
adi_: make sure you are setting LD_LIBRARY_PATH correctly; have you done that?
< moksh>
@zoq oh sorry, got confused.
< adi_>
how to set LD_LIBRARY_PATH correctly?
< rcurtin>
you should do 'export LD_LIBRARY_PATH=/path/to/libs/'
< rcurtin>
where /path/to/libs/ is the directory containing libmlpack.so
< adi_>
i have libmlpack.so in mlpack-2.2.5/build/lib/
< adi_>
there is a file named libmlpack.so in above path...
< rcurtin>
okay, did you do 'make install'?
< rcurtin>
if you did, look to see if it is in /usr/local/lib/ because that would be a typical place for it to be installed
< rcurtin>
then if it is there you can do `export LD_LIBRARY_PATH=/usr/local/lib/`
< adi_>
yes sir i did..
< adi_>
it worked sir...I did mlpack_logistic_regression -h in that folder...
< adi_>
so do i hv to put run those commands in that folder?
< adi_>
sir but...sinha@sinha-Inspiron-14-3467:/usr/local/lib$ mlpack_logistic_regression -t dataset.csv -v [WARN ] --output_model_file not given; trained model will not be saved. [FATAL] Cannot open file 'dataset.csv'. terminate called after throwing an instance of 'std::runtime_error' what(): fatal error; see Log::Fatal output Aborted (core dumped)
< adi_>
that command of dataset.csv is not being able to run successfully...
< moksh>
@zoq: I am getting this error while building with -DDEBUG=ON
< moksh>
Error: expected constructor, destructor, or type conversion before ‘(’ token math::RandomSeed(std::time(NULL));
robertohueso has quit [Quit: leaving]
< rcurtin>
adi_: did you try reading the error? dataset.csv does not exist so you will need to find a file to run the program with
< adi_>
yes sir i tried iris.csv but same result..
< adi_>
but mlpack_logistic_regression -h is giving output
< rcurtin>
okay, does iris.csv exist in the directory you are running the program in?
< rcurtin>
it may be worthwhile to spend some more time reading general command line documentation for linux, if you are not fully comfortable with the command line
sumedhghaisas2 has quit [Read error: Connection reset by peer]
sumedhghaisas has joined #mlpack
< zoq>
moksh: hm, strange, did you remove the build directory?
< moksh>
@zoq: No
< zoq>
moksh: Not sure, maybe that solves the issue
< moksh>
@zoq: Okay. I'll try and get back to you. Thanks
< adi_>
@rcurtin that iris.csv file doesnt exist in the usr/local/lib/ folder
< adi_>
and i then copied iris.csv file to that folder but that code doesnt seems to run!
< zoq>
adi_: You don't have to copy the file into the folder, you can just pass that path.
< zoq>
adi_: Also as rcurtin pointed out, we need some more information to help you.
< adi_>
sir i run the command in usr/local/lib folder..
< adi_>
actually mlpack-2.2.5 folder exists in Downloads..
< adi_>
i did make install
< rcurtin>
adi_: I really think, if you are not comfortable with using the command line and aren't familiar with the concepts of working directories, etc., it might be better for you to do reading first instead of having us help you with each problem
< rcurtin>
that way, you can probably get to the answer yourself and develop a greater understanding
< adi_>
ok sir..
< rcurtin>
in any case, make sure that the filename you pass to '-t' is either a fully specified path, or is in the current working directory
< zoq>
Agreed, there are tons of good tutorials out there.
< adi_>
ok sir..thanks..actually i already learned about basic linux commands ...but i'll do it ...
< moksh>
@zoq still getting the same error
< zoq>
moksh: What if you remove the line? Does this solve the issue?
< moksh>
Yes, it builds properly if I remove the line
< zoq>
and all you did was to add "mlpack::math::RandomSeed(std::time(NULL))" inside the test case right?
< moksh>
@zoq: Oh wait. I added it before the test case, outside it. Maybe that's the problem?
< zoq>
Ah, I think it is :)
< moksh>
@zoq: I thought before the test case meant outside it, sorry about that
< zoq>
No worries.
< adi_>
yippee it worked sir...Output is -> [INFO ] Loading '/home/sinha/Downloads/mlpack-2.2.5/build/iris.csv' as CSV data. Size is 4 x 150.
< adi_>
sir its 1 am hr...can i join u tomorrow....I m sorry for previous questions..i was rushing..
< moksh>
@zoq: That fixed it. Thanks. I'll let the build finish and get the output.
< moksh>
@zoq: Also can you please tell me some reasons why the test might have failed? As rcurtin mentioned in the ticket the tests are probabilistic so they might be unstable, but I can't properly understand why.
sumedhghaisas2 has joined #mlpack
sumedhghaisas has quit [Ping timeout: 256 seconds]
< zoq>
You start with a random function paramater initalization, so each run is different.
sumedhghaisas2 has quit [Ping timeout: 256 seconds]
< adi_>
Good night.
adi_ has quit [Quit: Page closed]
< moksh>
@zoq: Yeah, I understood that, but any reason why one particular optimizer might have failed? Would the initialization parameters of the optimizer have something to do with that?
< zoq>
moksh: It might be possible that it dosn't converge in the specified number of iterations, so oen easy solution is to increase the number of iterations, but we like to keep the test time low; so another idea is to increase/decrease the step size.
sumedhghaisas has joined #mlpack
< moksh>
@zoq: Oh okay. So I'll try running the test built with debug on and if the test fails consistently I'll try tuning the step size.
< moksh>
@zoq: I found the error was in one of the tests and just increasing the step size seems to have fixed it. Thanks again :) I'll push the change now
moksh has quit [Ping timeout: 260 seconds]
sumedhghaisas2 has joined #mlpack
sumedhghaisas has quit [Ping timeout: 256 seconds]
ImQ009 has quit [Read error: Connection reset by peer]