cameron.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/
karthikabinav has joined #mlpack
jbc_ has quit [Quit: jbc_]
karthikabinav has quit []
KristianHL has joined #mlpack
KristianHL has quit [Client Quit]
tarun_ml_enthusi has joined #mlpack
< tarun_ml_enthusi> hello..ive faced a problem i need help with
< tarun_ml_enthusi> while running simple programs with involving arma such as load a matrix, transpose,and write to a file...evertything worked smoothy
< tarun_ml_enthusi> but when i try running a command like sudo linear_regression --input_file testSet.csv -v
< tarun_ml_enthusi> it gives me this error
< tarun_ml_enthusi> Loading 'testSet.csv' as CSV data. Size is 5 x 7. error: qr(): use of LAPACK needs to be enabled terminate called after throwing an instance of 'std::logic_error' what(): qr(): use of LAPACK needs to be enabled
< tarun_ml_enthusi> i have checked in armadillo bits /config.hpp and these lines exist
< tarun_ml_enthusi> #if !defined(ARMA_USE_LAPACK) #define ARMA_USE_LAPACK
< tarun_ml_enthusi> do does anyone know what i should do ti resolve this error?
jbc_ has joined #mlpack
< naywhayare> tarun_ml_enthusi: do you have multiple versions of Armadillo installed?
< naywhayare> and can you tell me how you installed armadillo?
< tarun_ml_enthusi> nope only one version
< tarun_ml_enthusi> i downloaded it from theyre site and extracted it
< tarun_ml_enthusi> and then i put it in the include folder
tarun_ml_enthusi has quit [Ping timeout: 246 seconds]
tarun_ml_enthusi has joined #mlpack
witness___ has joined #mlpack
< naywhayare> tarun_ml_enthusi: so it sounds like you just downloaded armadillo and didn't compile it
< naywhayare> if I understood that correctly, you should go to the folder you extracted armadillo to and run ./configure, make, then make install
< tarun_ml_enthusi> but then how are the normal programs working?
< naywhayare> they don't use the LAPACK/BLAS functionality of armadillo
< naywhayare> so... armadillo is a header-only library, which means libarmadillo.so isn't strictly necessary
< tarun_ml_enthusi> oh okay so for that i need to run ./configue?
< tarun_ml_enthusi> configure*
< naywhayare> but armadillo also offers a runtime library which simply wraps LAPACK, BLAS, and other dependencies, so you can link as -larmadillo and not -llapack -lblas -larpack -lwhatever
< naywhayare> it's kind of a complex situation...
< naywhayare> yeah, ./configure then make should work
< tarun_ml_enthusi> so ok ill try it right now
< naywhayare> but if you are on ubuntu or any linux distro you can just use your package manager to install armadillo which is probably easier
< tarun_ml_enthusi> yeah using ubuntu
< naywhayare> ubuntu should be libarmadillo-dev or maybe libarmadillo4-dev
< naywhayare> yeah, just install that package instead
< tarun_ml_enthusi> gnome actually..doesnt matter tight
< tarun_ml_enthusi> right*
< naywhayare> it will get all the LAPACK and BLAS dependencies and whatnot, and the armadillo that gets installed will be configured correctly with LAPACK and BLAS so the more complex programs will work
< naywhayare> yeah, gnome doesn't matter
< tarun_ml_enthusi> ok so ill try using package manager to install
< naywhayare> yeah, much easier that way :)
< tarun_ml_enthusi> thanks :)
< tarun_ml_enthusi> and yeah i was looking into recommender systems and then ending up looking at matrix factorization and stuff...finally starting to understand something about that bug #372 :)
< naywhayare> matrix factorization and collaborative filtering are quite cool concepts
< naywhayare> heavily used in industry (as you might expect)
< tarun_ml_enthusi> yeah...i knew only the basic collaborative filtering before..they say the use of matrix factorization is relatively new
< tarun_ml_enthusi> there are so many methods of matrix factorization too
< naywhayare> yeah, very many
< naywhayare> we had some summer of code students last year who implemented many of these techniques
< naywhayare> NMF, SVD, batch SVD, incremental SVD, AMF, etc... lots of stuff
< tarun_ml_enthusi> ohh cool
< tarun_ml_enthusi> i have lots of catching up to do i guess :P
< naywhayare> :)
< naywhayare> there is tons to learn about machine learning... it can be overwhelming...
< tarun_ml_enthusi> hmm :)
< tarun_ml_enthusi> oh btw when i did configure it was saying LAPACK FOUND=NO
< tarun_ml_enthusi> do i need to download it separatly? and if yes how do i tell this think where to look?
< naywhayare> yeah, probably because you didn't have lapack installed
< naywhayare> you could download and install lapack by hand but it's much easier to just use the package manager... 'liblapack-dev' should be the package name and also 'libblas-dev'
< naywhayare> but if you installed libarmadillo-dev, it should get all of those dependencies automatically
< tarun_ml_enthusi> ohh...and where can libarmadillo_dev be found?
< naywhayare> in the repositories... just 'apt-get install libarmadillo-dev', or use your package manager or whatever
< naywhayare> it might be libarmadillo4-dev, I don't know
< tarun_ml_enthusi> oh ok thanks :) ....im new to linux ..still need some getting used to :P
< naywhayare> :)
< naywhayare> I have to get some lunch... I'll be back later
tarun_ml_enthusi has quit [Ping timeout: 246 seconds]