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/
kris1 has joined #mlpack
< sumedhghaisas> rcurtin: Hey Ryan... KRANN tests are failing for my PR... Don't know why. I did change anything there...
< sumedhghaisas> they are passing on AppVeyor
kris1 has quit [Quit: kris1]
< rcurtin> sumedhghaisas: the KRANNTest failures are known, don't worry about them
< rcurtin> there is an issue open for it: https://github.com/mlpack/mlpack/issues/734
kris1 has joined #mlpack
< kris1> Hi could you have a look at the serialize funciton
< kris1> in the RBM that i implemented t
< kris1> i can’t seem to make the test work for it
shikhar has joined #mlpack
< lozhnikov> Could you show the test?
< lozhnikov> 1. RbmXml, RbmText, RbmBinary shouldn't be initialized at all
< lozhnikov> 2. you don't serialize layers, therefore inSize and outSize aren't initialized
< kris1> so I would have to serialize the indivisual layers also if so i would have to call the serialise module from there class right
< lozhnikov> yeah, write something like
< lozhnikov> ar & visible;
< lozhnikov> ar & hidden;
< kris1> ar & data::CreateNVP(visible, "visible"); but isn’t doing the same thing
< kris1> I am not sure i haven’t read the boost seralisation docs
< lozhnikov> yes, I mean
< lozhnikov> ar & data::CreateNVP(visible, "visible");
< lozhnikov> ar & data::CreateNVP(hidden, "hidden");
< kris1> Well i had already done that in the Serialise function fo the RBM class
< lozhnikov> I didn't see that since you haven't push the recent version to git
< kris1> Ohhh i am sorry i will push them… also i think we would need the mnist dataset not only for rbm but also GAN and ssRBM
< kris1> So i think i would convert them to binary format and push them
< kris1> What do you think
< lozhnikov> I guess you should try some existing datasets first
< lozhnikov> I dont't think that uplodaing 11MB is a good idea
< kris1> I did try it with thyroid dataset but it did not work all other dataset’s are too small actually
< lozhnikov> try to extract some small part from the scikit example dataset
< kris1> Okay i will test it with that
< kris1> Also i had written the seralise function in rbm_impl.hpp file correctly in the previous commit as well
< kris1> I have not changed it in my present version
< kris1> so you can have a look there
< lozhnikov> In such away I have already said that you don't serialize layers
< kris1> So you are suggesting that i change ar & data::CreateNVP(visible, "visible"); —> with ar & visible
< kris1> Is that what you mena
< kris1> *mean
< lozhnikov> kris1: I don't see "ar & data::CreateNVP(visible, "visible");" in your implementation of Serialize() https://github.com/mlpack/mlpack/pull/1027/files
kris1 has quit [Quit: kris1]
kris1 has joined #mlpack
shikhar has quit [Quit: WeeChat 1.7]
mikeling has joined #mlpack
shikhar has joined #mlpack
travis-ci has joined #mlpack
< travis-ci> mlpack/mlpack#2684 (master - 483833e : Ryan Curtin): The build is still failing.
travis-ci has left #mlpack []
shikhar has quit [Quit: WeeChat 1.7]
govg has quit [Ping timeout: 255 seconds]
mentekid has quit [Quit: Leaving.]
mentekid has joined #mlpack
chenzhe has joined #mlpack
chenzhe has quit [Remote host closed the connection]
chenzhe has joined #mlpack
chenzhe has quit [Quit: chenzhe]
chenzhe has joined #mlpack
partobs-mdp has joined #mlpack
partobs-mdp has quit [Client Quit]
< chenzhe> This might be an easy question: how do we check arma::mat values in debugger like gdb or lldb?
< chenzhe> I am using clion, it shows me mat variable has component: "mem" and "mem_local"
< chenzhe> mem is of const double *const
< chenzhe> and mem_local is of double[16], the dimension looks not the same as my matrix
< rcurtin> chenzhe: yeah, you want the mem[] array
< rcurtin> but to get an individual element you will have to offset it yourself a little bit
< rcurtin> since armadillo matrices are column-major (i.e. columns are contiguous), then to get the memory location of location (r, c) will be c * n_rows + r
< rcurtin> so, i.e., position (2, 3) -> mem[3 * n_rows + 2]
< rcurtin> the mem_local[] array is, I think, what is used for fixed-size matrices
< chenzhe> rcurtin: Thanks! ^_^
chenzhe1 has joined #mlpack
< rcurtin> sure, let me know if I can help with anything else :)
chenzhe has quit [Ping timeout: 255 seconds]
chenzhe1 is now known as chenzhe
chenzhe has quit [Quit: chenzhe]
mikeling has quit [Quit: Connection closed for inactivity]
< kris1> appvoyer and travis are failing arma::mat X; X = {{1, 0, 1},{0,0,0}}; but this passes locally
< kris1> any ideas why this is happening
< kris1> thanks mikhail just saw your comment
< lozhnikov> not sure, but maybe Travis's version of armadillo doesn't support that
chenzhe has joined #mlpack
mentekid has quit [Quit: Leaving.]
sikoizon has joined #mlpack
< sikoizon> while downloading version 2.2.3 of mlpack on linux 64 machine with ubuntu 16.04 LTS
< sikoizon> and using the tutorial for build on mlpack .. i am getting an error
< sikoizon> help
< sikoizon> this is the page i referred to
< sikoizon> <rcurtin>
< sikoizon> will see the log of mlpack for irc .. answer for my question .. thanks
sikoizon has quit [Quit: Page closed]