--- Log opened Thu Dec 22 00:00:19 2011 | ||
-!- blackburn [~blackburn@188.168.2.99] has quit [Quit: Leaving.] | 01:46 | |
-!- Ram108 [~amma@14.96.61.57] has joined #shogun | 06:57 | |
-!- Ram108 [~amma@14.96.61.57] has quit [Ping timeout: 244 seconds] | 09:49 | |
-!- Ram108 [~amma@14.96.183.218] has joined #shogun | 10:09 | |
-!- axitkhurana [~akshit@14.98.107.235] has joined #shogun | 11:28 | |
-!- axitkhurana [~akshit@14.98.107.235] has quit [Quit: Leaving.] | 11:40 | |
-!- Ram108 [~amma@14.96.183.218] has quit [Ping timeout: 255 seconds] | 12:50 | |
-!- puneetgoyal [~puneetgoy@117.197.162.106] has joined #shogun | 13:47 | |
puneetgoyal | hey, what can I check in the attachment , if it can contribute to probability of the mail being a spam or a ham | 13:53 |
---|---|---|
-!- heiko [~heiko@p4FED50C1.dip.t-dialin.net] has joined #shogun | 15:16 | |
heiko | sonney2k, around? | 15:16 |
sonne-work | yes | 15:21 |
sonne-work | heiko: ^ | 15:21 |
heiko | sonney2k, finally we meet :) | 15:21 |
sonne-work | yes, took a while... seems like summer aka spare time is over :) | 15:22 |
heiko | indeed :) | 15:22 |
heiko | how is it going? | 15:22 |
sonne-work | so how is it going? do you have some time now over the years? | 15:22 |
sonne-work | I am basically busy all the time nowadays | 15:22 |
sonne-work | but I try to do small changes whenever possible | 15:23 |
heiko | well same here, the UCL master is like crazy workload wise | 15:23 |
heiko | but I have holidays until mid january now | 15:23 |
heiko | I visit family and friends | 15:23 |
heiko | but have some time in between | 15:23 |
heiko | so it should be possible to do something in these weeks now | 15:24 |
sonne-work | would be nice | 15:24 |
sonne-work | I don't know if you received my email from today but maybe you explain a bit more what you have in mind | 15:24 |
heiko | yes i did | 15:24 |
heiko | well basically the most annoying problem is to allocate an "empty" parameter | 15:25 |
sonne-work | not sure what that means but go on | 15:26 |
heiko | so | 15:26 |
heiko | currently, if you want to load a parameter from a file | 15:26 |
heiko | you call load on the instance of the object you want to load | 15:26 |
heiko | but all the memory is already allocated | 15:26 |
heiko | all variables are registered etc | 15:26 |
heiko | but now imagine you dont have a class instance and want to load something from memory | 15:26 |
heiko | imageine a class has an SGVector as a variable | 15:27 |
heiko | these normally are member variables | 15:27 |
heiko | so you have to allocate a SGVector | 15:27 |
heiko | and then also the actual memory in that sgvector | 15:27 |
heiko | know what I mean? | 15:28 |
heiko | It is quite hard to load data from file when you dont got the class instance to load into, in summary | 15:28 |
heiko | the methods new_cont/delete_cont | 15:28 |
heiko | do not work in this case | 15:28 |
sonne-work | why don't you add load/save methods to SGVector? | 15:29 |
sonne-work | then you could just say vec.load etc | 15:29 |
sonne-work | One more thing I don't understand - when does it happen that you don't have a class instance and want to load sth? | 15:31 |
heiko | This happens when you want to migrate parameters | 15:32 |
heiko | you got a class with a parameter, but the one in the file is different | 15:32 |
heiko | so you cannot load directly | 15:32 |
heiko | but to migrate, you have to load somehow | 15:32 |
heiko | Currently, when you got a sgvector | 15:33 |
heiko | you have to | 15:33 |
sonne-work | so you can load only basic types right? | 15:33 |
heiko | yes | 15:33 |
sonne-work | like scalars, or double*, int / SGVector<double> | 15:33 |
sonne-work | each of them need extra treatment though | 15:34 |
heiko | data=SG_MALLOC(SGMatrix<char>, 1); | 15:34 |
sonne-work | no code reuse | 15:34 |
heiko | yes | 15:34 |
sonne-work | what is that data= line? | 15:34 |
heiko | to be able to load data from a file you need a TParameter whichs parameters point to already allocated data | 15:35 |
heiko | in case of sgmatrix, that sgmatrix is a class variable | 15:35 |
heiko | thats why you have to do this line before loading | 15:35 |
heiko | if you dont, new_data/delete_data of TParameter fails | 15:35 |
heiko | but this gets completely messy if more types come in | 15:36 |
sonne-work | what other types? | 15:36 |
heiko | sgvector | 15:36 |
heiko | vector | 15:36 |
heiko | scalar | 15:36 |
heiko | everytime this happens, a case distinction is needed, for all types. | 15:36 |
sonne-work | yes | 15:36 |
sonne-work | but I don't see a way around this | 15:37 |
sonne-work | do you? | 15:37 |
heiko | it would be nice to have a method that creates a TParameter with all corresponding data allocated | 15:37 |
sonne-work | I understand that SGObject based classes can be created via new_sgserializable | 15:38 |
heiko | ? | 15:39 |
sonne-work | maybe you could do the same thing for SGVector / SGMatrix / SGSparseMatrix / SGStringList | 15:39 |
sonne-work | CSGObject* new_sgserializable(const char* sgserializable_name, | 15:40 |
sonne-work | EPrimitiveType generic); | 15:40 |
sonne-work | shogun/base/class_list.h | 15:40 |
sonne-work | scalars still need extra treatment though | 15:41 |
heiko | mmh | 15:42 |
sonne-work | any other ideas? | 15:43 |
heiko | I thought of what would be if this SGVector stuff would be done differently: | 15:44 |
heiko | I find it very confusing that when registering the variables | 15:44 |
heiko | that the adress of SGVector/SGMatrix are registered | 15:45 |
heiko | why not only register the array and the length | 15:45 |
heiko | then the wrapper stuff would still be ok. | 15:45 |
heiko | but the problems with allocating SGVector structs would not be there | 15:46 |
heiko | then the new_cont / delete_cont methods would only be operating on arrays | 15:46 |
heiko | which would simplify things enoumously | 15:46 |
heiko | then allocating empty structures is also more easy | 15:46 |
heiko | Perhaps I am wrong with this | 15:46 |
sonne-work | so what you are saying is that you would in the add(SGVector) method of parameter register the ptrs? | 15:47 |
heiko | yes, just register the array and the length of the SGVector | 15:47 |
heiko | then it would be completely out of sight for the TParameter class | 15:47 |
sonne-work | yes that would work | 15:48 |
heiko | also then these things would not be appearing anymore: | 15:48 |
heiko | data_start=&(((SGVector<char>*)m_parameter)->vector); | 15:48 |
sonne-work | which is highly illegal anyways | 15:48 |
heiko | (in new_cont) | 15:48 |
heiko | yes this is dirty to the max | 15:48 |
heiko | it is there because of the problem to get the array of a sgvector you dont know the type of | 15:49 |
sonne-work | so to summarize: one would have the very basic types like, double, int, ..., then vector like things double*, int / and matrix/string/sparse matrix and SGObject | 15:49 |
sonne-work | and we map SGVector etc to vector etc | 15:50 |
heiko | yes | 15:50 |
sonne-work | so we have the more basic representation underneath | 15:50 |
heiko | true | 15:50 |
sonne-work | certainly one way to do it | 15:50 |
heiko | When using SGVector underneath, a lot of stuff has to be changed in the new_cont methods | 15:51 |
sonne-work | the other would be to remove all this double*, int serialization stuff and replace it with SGVector stuff | 15:51 |
sonne-work | yes | 15:51 |
heiko | because everything in the low level parameter stuff is based to work on only arrays | 15:51 |
sonne-work | which would totally go away | 15:51 |
heiko | yes | 15:51 |
sonne-work | I mean you could work with references only | 15:51 |
sonne-work | no longer any pointers but just the object itself | 15:52 |
heiko | yes, possible | 15:52 |
heiko | but this means to change it at all places at once | 15:52 |
sonne-work | but | 15:52 |
sonne-work | no reference is possible with void | 15:52 |
sonne-work | hah | 15:52 |
sonne-work | so it would need extra treatment... | 15:53 |
sonne-work | I certainly find the other alternative easier | 15:53 |
sonne-work | so I would rather go for that one for now | 15:53 |
heiko | this changing the new_cont methods all over is my horror when continuing with the current approach | 15:53 |
sonne-work | I understand that | 15:53 |
sonne-work | if we have had SGVector etc from the beginning things would not look so bad :) | 15:54 |
heiko | yes, but this is always the case when developing software :D | 15:54 |
sonne-work | anyway, it also makes sense to still have the legacy double*, int serialization | 15:54 |
heiko | but this is easily possible isnt it? | 15:55 |
heiko | the register methods still take arrays as parameter | 15:55 |
sonne-work | well that still works as usual - so yes | 15:55 |
heiko | so the sgvector register methods are only wrappers for these | 15:55 |
sonne-work | I only meant if one would totally throw that away and convert to the SGVector etc stuff | 15:55 |
heiko | ok | 15:56 |
sonne-work | the only problem with the double* etc approach is that you need to ensure all variables in sgvector are serialized | 15:56 |
sonne-work | hah | 15:56 |
sonne-work | maybe you add some other register_serialization_varaibles() in sgvector itself | 15:57 |
sonne-work | which is then called when you call add(SGVector) in Parameter | 15:57 |
heiko | oh yes, you mean this do_free stuff etc? | 15:57 |
sonne-work | yeah | 15:57 |
heiko | ok possible, but how to deserialise? | 15:58 |
sonne-work | in the same way | 15:59 |
sonne-work | you have your SGObject where a SGVector vec lives | 15:59 |
heiko | also there might be problems with the names of the sub-variables in SGVector | 15:59 |
sonne-work | it is registering its parameters | 15:59 |
sonne-work | and you fill in the values | 15:59 |
sonne-work | problems with the names? | 16:00 |
sonne-work | I guess devil is in the details again | 16:00 |
heiko | nevermind the names, just thinking ... | 16:01 |
heiko | oif problems | 16:01 |
heiko | I want to think of the migration to avoid more problems when doing this | 16:01 |
heiko | still, how to load from a file without a instance | 16:02 |
heiko | when there are only arrays, its easy | 16:02 |
heiko | the boolean would be a normal variable | 16:03 |
sonne-work | but you have that problem no matter where you register the parameter right ? | 16:03 |
sonne-work | yes | 16:04 |
heiko | which problem? | 16:04 |
sonne-work | loading w/o object instance | 16:04 |
sonne-work | I mean ok, you can load the varialbes into memory | 16:04 |
sonne-work | and then do all the migration | 16:04 |
sonne-work | and then later on - how does it get into the final object / SGVector? | 16:05 |
sonne-work | ahh no I see | 16:05 |
sonne-work | no problem nevermind | 16:05 |
heiko | after the migration is complete the migrated data is copied into the actual registered array of the SGVector | 16:06 |
heiko | and also the length | 16:06 |
sonne-work | yes exactly | 16:06 |
heiko | then the migration does not even see these wrapper types for vectors and matrices | 16:06 |
sonne-work | so I don't see a problem w/ migration for now | 16:06 |
sonne-work | yes | 16:06 |
sonne-work | so you could give this a try | 16:07 |
heiko | One problem might occur. people who have stored SGVectors in files would not be able to load anymore | 16:07 |
sonne-work | but you could map that back to the old vector way or? | 16:07 |
heiko | oh, theoretically .. yes | 16:08 |
sonne-work | case CT_VECTOR: case CT_SGVECTOR: | 16:09 |
sonne-work | actually - would it even be any work? | 16:09 |
heiko | dont really know, but I changed some stuff to save/load SGVectors, they are handled a bit differently, will have a look | 16:09 |
sonne-work | it looks like you treat them the same way like vectors and then only determined data start etc | 16:10 |
sonne-work | and fill things in | 16:10 |
sonne-work | so it seems like no problem | 16:10 |
sonne-work | anyways - I would give this a go | 16:10 |
sonne-work | so if you have time - it would be great if you could first cleanup Parameter* | 16:11 |
sonne-work | and then continue with the migration | 16:11 |
heiko | yes, this is the way | 16:11 |
sonne-work | I will try to get buildbots to work more reliably | 16:11 |
sonne-work | but it is tough | 16:11 |
sonne-work | (but we have one for cygwin and for linux now) | 16:11 |
sonne-work | osx is coming | 16:12 |
heiko | basically, migration works, the SGVector stuff was the thing that caused all these problems that I coulnt solve in the summer | 16:12 |
heiko | ah cool | 16:12 |
heiko | cygwin? :) | 16:12 |
sonne-work | sounds great | 16:12 |
sonne-work | windows 7 | 16:12 |
heiko | really cool | 16:12 |
sonne-work | and cygwin running there | 16:12 |
heiko | I can still hear my old professor complaining that my bachelor stuff does not run on windows | 16:12 |
heiko | (what they were using) | 16:12 |
sonne-work | when we have these build bots running more reliably we can even create binary release-snapshots | 16:13 |
heiko | that would also be great, especially for cygwin | 16:13 |
heiko | well, ok will be off now for some sport, see you later, Ill inform you about the progress | 16:14 |
sonne-work | thanks for continuing to work on this | 16:14 |
sonne-work | would be very cool to have this feature! | 16:14 |
sonne-work | see you | 16:14 |
heiko | sonne-work, I just looked at Parameter.h | 16:19 |
heiko | what we talked about would mean to change all the add* methods from using pointers to SG* to using copy-by-value right? | 16:20 |
heiko | So void add(SGVector<bool>* param, const char* name,const char* description=""); | 16:20 |
heiko | gets | 16:20 |
heiko | add(SGVector<bool> param, const char* name,const char* description=""); | 16:20 |
heiko | for all SGVector/SGMatrix entries, right? | 16:21 |
heiko | what about the SGSparseVector SGString stuff? Is this also afected? | 16:24 |
heiko | ah no...the length needs to be a the real thing, not a copy, sorry | 16:29 |
sonne-work | yes everything is affected | 16:32 |
sonne-work | we have SGStringList | 16:32 |
sonne-work | SGSparseMatrix | 16:32 |
sonne-work | SGMatrix | 16:32 |
sonne-work | SGVector | 16:32 |
sonne-work | all by reference yes | 16:32 |
sonne-work | all these classes will copy their local variables in the copy constructor | 16:33 |
sonne-work | so SGVector x = SGVector(foo, foo_len) | 16:33 |
sonne-work | will have x.vector == fo and x.vlen==foo_len | 16:33 |
sonne-work | heiko: btw one more thing - I am trying to write some print function to improve user experience from python_modular | 16:34 |
sonne-work | that needs the list of parameters - but you return SGVector<char*> for that | 16:35 |
sonne-work | but it would be much better to use SGStringList<char> for this | 16:35 |
heiko | ehm | 16:35 |
sonne-work | I changed the code already | 16:35 |
sonne-work | but haven't committed yet | 16:35 |
sonne-work | heiko: ahh and btw you can write | 16:35 |
heiko | where is the method again? | 16:35 |
sonne-work | SGVector<int>(x, x_len); | 16:36 |
sonne-work | err | 16:36 |
sonne-work | SGVector<int> my_vector(x, x_len); | 16:36 |
sonne-work | to declare my_vector | 16:36 |
heiko | instead of? | 16:36 |
sonne-work | SGVector<int> my_vector = SGVector<int>(x, x_len); | 16:36 |
heiko | ah ok, thx | 16:37 |
heiko | modshogun_wrap.cxx: In Funktion >>void shogun_CSGObject___setstate__(shogun::CSGObject*, PyObject*)<<: | 16:41 |
heiko | modshogun_wrap.cxx:5722:13: Fehler: expected >>;<< before >>fstream<< | 16:41 |
heiko | make[1]: *** [modshogun_wrap.cxx.o] Fehler 1 | 16:41 |
heiko | when compiling python_modular | 16:41 |
sonne-work | trying to reproduce this | 16:47 |
sonne-work | did you do a git clean -dfx before building? (will delete all files not in git) | 16:47 |
heiko | rebuilding ... | 16:55 |
heiko | yes happens | 16:56 |
heiko | still | 16:56 |
-!- Ram108 [~amma@59.161.189.110] has joined #shogun | 17:05 | |
sonne-work | I just rebuild here | 17:15 |
sonne-work | doesn't happen | 17:15 |
sonne-work | did you do make install etc? | 17:16 |
sonne-work | heiko: please show me the offending line | 17:16 |
heiko | fstream = new CSerializableAsciiFile(fname, 'r'); | 17:18 |
heiko | comes after: | 17:18 |
heiko | if (!pickle_ascii) | 17:18 |
heiko | sg_io->message(MSG_ERROR,"SGBase.i", 346, "File contains an HDF5 stream but " \ | 17:18 |
heiko | "Shogun was not compiled with HDF5" \ | 17:18 |
heiko | " support! - cannot load.") | 17:18 |
sonne-work | ahh so a problem w/ hdf5 | 17:21 |
sonne-work | let me check | 17:21 |
sonne-work | but that is ok | 17:22 |
sonne-work | you just don't have hdf5 installed | 17:22 |
sonne-work | heiko: how does this relate to the error above? | 17:23 |
heiko | I thought because of the missing ; before fstream | 17:24 |
heiko | because: modshogun_wrap.cxx:5722:13: Fehler: expected >>;<< before >>fstream<< | 17:24 |
sonne-work | heiko: I meant just open modshogun_wrap.cxx in an editor | 17:33 |
sonne-work | and then show me the surrounding lines | 17:33 |
heiko | thats what I did | 17:33 |
heiko | fstream = new CSerializableAsciiFile(fname, 'r'); | 17:33 |
heiko | is the line | 17:33 |
heiko | and the other stuff comes before | 17:34 |
sonne-work | and before that? | 17:34 |
sonne-work | hmm | 17:34 |
sonne-work | ahh after the print | 17:34 |
sonne-work | fixed | 17:35 |
CIA-1 | shogun: Soeren Sonnenburg master * r5a8c07d / src/interfaces/modular/SGBase.i : add missing ';' - http://git.io/Z7HAIw | 17:35 |
sonne-work | " support! - cannot load."); instead of " support! - cannot load.") | 17:35 |
heiko | kk :) | 17:39 |
shogun-buildbot | build #381 of python_static is complete: Failure [failed configure] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_static/builds/381 blamelist: sonne@debian.org | 17:39 |
-!- blackburn [~blackburn@188.168.4.33] has joined #shogun | 17:42 | |
heiko | compiles for me now | 17:43 |
blackburn | hey what's up | 17:43 |
heiko | hey blackburn, sorry I was ust about to go :) | 17:46 |
blackburn | heiko: when? | 17:46 |
heiko | now | 17:46 |
heiko | sorry, i have an appointment | 17:47 |
heiko | will talk to you later :) bye | 17:47 |
blackburn | ah | 17:47 |
blackburn | sure | 17:47 |
blackburn | bye | 17:47 |
Ram108 | hi | 17:47 |
blackburn | Ram108: hi | 17:48 |
blackburn | Ram108: will you be available in next hour? | 17:48 |
Ram108 | blackburn: could u help me out with the rest of it now that my exams are over and i am free :) | 17:48 |
Ram108 | yeah i ll be :) | 17:49 |
blackburn | Ram108: yeah sure but not now | 17:49 |
Ram108 | sure take ur time :) | 17:49 |
blackburn | just came home and now off for some dinner :) | 17:49 |
Ram108 | definitely wen u r free :) | 17:49 |
Ram108 | take ur time :) | 17:49 |
blackburn | ok, I'll write to you soon | 17:49 |
Ram108 | :) | 17:49 |
-!- heiko [~heiko@p4FED50C1.dip.t-dialin.net] has left #shogun [] | 17:52 | |
-!- ishaanmlhtr [~ishaan@115.240.93.129] has joined #shogun | 18:07 | |
blackburn | Ram108: re | 18:24 |
blackburn | Ram108:how can I help you? | 18:25 |
-!- Ram108 [~amma@59.161.189.110] has quit [Ping timeout: 244 seconds] | 18:26 | |
-!- Ram108 [~amma@14.96.20.215] has joined #shogun | 18:35 | |
blackburn | sonne-work: sonney2k: any of sonnes around? ;) | 18:37 |
blackburn | Ram108: I'm available | 18:37 |
Ram108 | lol :) | 18:37 |
Ram108 | thank u :) | 18:37 |
Ram108 | hmmm can we get started? | 18:38 |
blackburn | Ram108: yeah | 18:39 |
Ram108 | i have made the realfeatures() | 18:39 |
Ram108 | now i have to get make the test and train vectors | 18:40 |
Ram108 | right? | 18:40 |
blackburn | Ram108: ehmm one real features for test and one for train | 18:41 |
Ram108 | okay.........meaning after i create the test and train arrays i ll have to call the realfeatures shogun method and pass them to it right? | 18:43 |
blackburn | yes, it is just object representing data | 18:46 |
Ram108 | k am working on it now........ am using all the data for train and just about 5 of them for test......... | 18:47 |
blackburn | sure | 18:49 |
Ram108 | :) | 18:49 |
puneetgoyal | hey, is there any function to remove the html tags? | 19:26 |
blackburn | puneetgoyal: ehm from? | 19:55 |
Ram108 | it says train or test features dimension mismatch | 19:56 |
Ram108 | what do i have to do? | 19:56 |
blackburn | Ram108: I guess you have to transpose matrix before creating realfeatures | 19:57 |
Ram108 | oh i have used all the data u gave on that webpage for test and for train i have spooled out some of them (about 5) and used it | 19:58 |
Ram108 | now transpose both and feed it to the realfeatures() method right? | 19:58 |
blackburn | yes | 19:59 |
Ram108 | k | 20:00 |
blackburn | feature matrix should have dim rows | 20:00 |
blackburn | and N cols | 20:00 |
puneetgoyal | blackburn: sry, I thought my problem was in removing the html tags...but it was not.....I was removing the stop words from the payload of an email, Its working...but not correctly...I mean most of the stop words are removed but not all of them | 20:00 |
blackburn | puneetgoyal: I see, just try to find some python lib for that | 20:01 |
puneetgoyal | I have searched a lot of them....its working but dont know why some problem is coming...like one of the stopword is 'the' ...its removing many of its instances in the payload...but not all of them | 20:02 |
Ram108 | blackburn: same error again | 20:05 |
blackburn | Ram108: check feature matrices sizes | 20:10 |
blackburn | both should have 4 rows | 20:10 |
Ram108 | yeah both does | 20:12 |
Ram108 | i ll use the same matrix for both test and train | 20:13 |
Ram108 | i ll see if that works | 20:13 |
Ram108 | this is the error i am getting: | 20:17 |
Ram108 | File "/usr/local/lib/python2.6/dist-packages/modshogun.py", line 20351, in train | 20:17 |
Ram108 | return _modshogun.Machine_train(self, data) | 20:17 |
Ram108 | SystemError: [ERROR] label[-1717986918]=5.87727e-270 is not an integer | 20:17 |
blackburn | yeah something is wrong with labels you use | 20:17 |
-!- heiko [~heiko@p4FED50C1.dip.t-dialin.net] has joined #shogun | 20:20 | |
heiko | blackburn, around? | 20:20 |
blackburn | heiko: yeah | 20:20 |
heiko | hi | 20:20 |
heiko | how is it going? | 20:20 |
heiko | blackburn, I noticed that all the python_modular tests are failing, is that also true for you? | 20:21 |
blackburn | heiko: ALL? | 20:21 |
heiko | yes, memerror. | 20:21 |
blackburn | heiko: going fine, now some exams and other hard things with studies | 20:21 |
blackburn | and you? | 20:21 |
blackburn | memerror? | 20:21 |
blackburn | ehm | 20:21 |
heiko | At first I thought these were my changes | 20:21 |
blackburn | let me check | 20:21 |
heiko | but I reverted and they still fail | 20:21 |
heiko | I am on holidays :) relaxing at my parents and writing applications for PHD. | 20:22 |
blackburn | heiko: cool :) | 20:23 |
heiko | btw: I finally got a new computer with an SSD to have no more harddisk erors :) | 20:23 |
blackburn | hah | 20:25 |
blackburn | you would have some in 5-6 years I guess ;) | 20:25 |
heiko | yes, but i am using this trim stuff :) | 20:26 |
heiko | however, its fast | 20:26 |
heiko | awesomely fast | 20:26 |
heiko | btw shogun 1.1, also all python_modular tests fail (git branch) | 20:26 |
blackburn | ?!! | 20:26 |
blackburn | damn that's bad | 20:26 |
heiko | but different error | 20:26 |
heiko | perhaps its my installation or something | 20:27 |
blackburn | heiko: compiling now | 20:28 |
heiko | which one? | 20:28 |
blackburn | latest | 20:28 |
blackburn | heiko: btw we are now using git issues | 20:28 |
blackburn | https://github.com/shogun-toolbox/shogun/issues?sort=created&direction=desc&state=open | 20:28 |
heiko | yes I saw it, pretty cool | 20:29 |
blackburn | feel free to open new, comment, etc | 20:29 |
heiko | ok :) | 20:29 |
heiko | just had to implement gaussian processes for my studies int matlab :) | 20:29 |
blackburn | heiko: that would be awesome to have C++ one here | 20:29 |
heiko | if I only had more time ... :) | 20:30 |
heiko | aaah | 20:31 |
heiko | I did not checkout data :( | 20:32 |
heiko | that might be the reason for the error | 20:32 |
heiko | sorry | 20:32 |
blackburn | yay | 20:32 |
blackburn | Segmentation fault | 20:32 |
heiko | ah ok | 20:32 |
heiko | well, then its not my fault at least :) | 20:32 |
heiko | dinner is ready here, see you later! | 20:32 |
blackburn | heiko: ok, see you :) | 20:33 |
blackburn | I'm fucking shocked | 20:33 |
blackburn | sonney2k: I guess you broke python modular | 20:33 |
blackburn | totally | 20:33 |
blackburn | :D | 20:33 |
Ram108 | blackburn: can i email u my code? | 20:34 |
blackburn | Ram108: feel free | 20:34 |
Ram108 | blackburn: i feel lost........ did u feel the same way wen u started learning this feild? | 20:34 |
Ram108 | field* | 20:35 |
blackburn | Ram108: some kind of | 20:35 |
Ram108 | sheesh i really dont kno were to start............ am not able to compile a code........ let alone making 1........ lol | 20:36 |
-!- heiko [~heiko@p4FED50C1.dip.t-dialin.net] has quit [Read error: Operation timed out] | 20:36 | |
blackburn | Ram108: well I could write some snippet, but tomorrow | 20:37 |
Ram108 | sure..... thanks..... | 20:38 |
-!- ishaanmlhtr [~ishaan@115.240.93.129] has quit [Quit: Leaving] | 20:40 | |
Ram108 | well were do i really get started to learn machine learning? | 20:42 |
blackburn | Ram108: not sure I understood Q? | 20:43 |
Ram108 | never mind......... lol well probably if i could get ur snippet tomorrow i could compare it with mine and fix all the errors | 20:45 |
Ram108 | i guess its more or less the same for all the other classifiers right? | 20:45 |
blackburn | Ram108: yes, it doesn't matter that classifier do you use | 20:46 |
Ram108 | hmmm okay thanks........ :) | 20:46 |
blackburn | anyway in shogun you will have to use features,labels and etc | 20:46 |
Ram108 | okay | 20:47 |
Ram108 | okay meet u tomorrow then its getting late here......... lol :) | 20:48 |
Ram108 | gd night :) | 20:48 |
-!- Ram108 [~amma@14.96.20.215] has quit [Quit: Ex-Chat] | 20:50 | |
-!- blackburn [~blackburn@188.168.4.33] has quit [Ping timeout: 252 seconds] | 20:53 | |
-!- blackburn [~blackburn@188.168.3.26] has joined #shogun | 20:56 | |
-!- blackburn [~blackburn@188.168.3.26] has quit [Ping timeout: 252 seconds] | 21:00 | |
-!- blackburn [~blackburn@188.168.4.242] has joined #shogun | 21:01 | |
puneetgoyal | blackburn: you there? | 21:22 |
blackburn | puneetgoyal: yeap | 21:22 |
puneetgoyal | blackburn: you asked me to compare the array from a test mail with some other array | 21:23 |
puneetgoyal | ['this','is','spam'] is 1.0 to ['this','is','spam'], but 0.6667 to ['this','is','sparta'] | 21:23 |
puneetgoyal | but, how do I find the list with which I have to compare it to | 21:23 |
puneetgoyal | oh sry, I was having wrong results because I took only a little training emails...now I am getting somewhat good results | 21:27 |
-!- blackburn [~blackburn@188.168.4.242] has quit [Ping timeout: 240 seconds] | 21:27 | |
-!- blackburn [~blackburn@188.168.4.225] has joined #shogun | 21:28 | |
blackburn | puneetgoyal: sorry bad connection | 21:29 |
puneetgoyal | blackburn: no problem, Now I am done building a dictionary counting the no. of words....using which I trained one dictionary of words...Now I am trying to compare a test dict with the trained one | 21:30 |
@sonney2k | blackburn, what happens w/ python modular? | 22:43 |
@sonney2k | what is broken? | 22:43 |
@sonney2k | ahh you mean the examples are failing | 22:44 |
@sonney2k | yes sure I now return some short representation upon str instead of the serialized stuff | 22:45 |
@sonney2k | so that is ok | 22:45 |
blackburn | sonney2k: yes, I'm talking about python_modular examples | 22:47 |
CIA-1 | shogun: Soeren Sonnenburg master * r38cbdc3 / (src/shogun/base/SGObject.cpp src/shogun/base/SGObject.h): use SGStringList<char> instead of SGVector<char*> - http://git.io/pNw3WQ | 22:49 |
shogun-buildbot | build #390 of r_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_static/builds/390 blamelist: sonne@debian.org | 22:58 |
shogun-buildbot | build #394 of octave_static is complete: Failure [failed test_1] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/394 blamelist: sonne@debian.org | 22:58 |
-!- heiko [~heiko@p4FED50C1.dip.t-dialin.net] has joined #shogun | 23:15 | |
CIA-1 | shogun: Soeren Sonnenburg master * rcc60a77 / (5 files in 3 dirs): | 23:43 |
CIA-1 | shogun: Merge pull request #343 from karlnapf/master | 23:43 |
CIA-1 | shogun: changed internal storage of SGVector - http://git.io/Le6GCw | 23:43 |
CIA-1 | shogun: Heiko Strathmann master * r2b685fe / examples/undocumented/libshogun/serialization_file_formats.cpp : fixed a typo - http://git.io/9Hoiyg | 23:43 |
shogun-buildbot | build #395 of octave_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_static/builds/395 | 23:54 |
shogun-buildbot | build #392 of r_static is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_static/builds/392 | 23:56 |
--- Log closed Fri Dec 23 00:00:19 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!