Open in new window / Try shogun cloud
--- Log opened Sun Dec 18 00:00:19 2011
-!- ishaanmlhtr [~ishaan@] has quit [Ping timeout: 252 seconds]01:34
@sonney2kblackburn, what is a data set downloader?01:42
blackburnsonney2k: we can't include say mnist01:42
blackburnbut can add some downloader01:42
blackburnscikits did it01:42
@sonney2kblackburn, I see - maybe we should upload all our datasets to and then just grab it from there01:43
@sonney2kthat would have the advantage that we can easily read it from python01:43
blackburnsonney2k: yes something like that01:43
blackburnsonney2k: why don't you sleeping? :)01:43
@sonney2kI could ask you the same thing...01:44
@sonney2kI am doing the new builtin serialization01:44
blackburnsonney2k: hah, I was arguing about google new face detection technology01:45
blackburnsonney2k: could you take a look on my paper?01:47
@sonney2ksend it to me01:48
@sonney2ktomorrow I have a bit of time01:48
blackburnsonney2k: ok thanks in advance01:48
blackburnit is 5 page so should not take a long ;)01:49
blackburn3 in fact, references and pics takes 201:49
-!- ishaanmlhtr [~ishaan@] has joined #shogun01:53
blackburnsonney2k: good night01:59
@sonney2ksleep well :)01:59
@sonney2kwell 4 pages is the limit02:00
@sonney2k+1 page citations02:00
blackburnsonney2k: is it a hard limit?02:00
blackburnsonney2k: how can it be done?..02:00
@sonney2kshogun had only 3 pages02:01
@sonney2k+1 ref02:01
@sonney2kso be happy that you have one more page02:01
blackburnsonney2k: will 4+1 work?02:01
blackburnI mean is it strict 4 including refs?02:01
@sonney2k4 strict for content02:03
@sonney2k+1 page refs02:03
blackburnI'm worried how to cut it in case of it is too large :)02:03
@sonney2kusually no problem02:04
blackburnsonney2k: okay thanks02:06
blackburnsleep well too, see yuo02:06
-!- blackburn [~blackburn@] has quit [Quit: Leaving.]02:06
-!- ishaanmlhtr [~ishaan@] has quit [Ping timeout: 255 seconds]04:03
-!- ishaanmlhtr [~ishaan@] has joined #shogun04:25
-!- ishaanmlhtr [~ishaan@] has quit [Quit: Leaving]04:28
-!- Ram108 [~amma@] has joined #shogun05:30
-!- ishaanmlhtr [~ishaan@] has joined #shogun05:38
ishaanmlhtrRam108:are you done with classifying the dataset that blackburn gave using kNN?05:39
Ram108nope sorry i had exams last week just got free now........... i ll be starting again today05:43
Ram108are u done with it ishaan?05:44
ishaanmlhtrno, i am stuck up.getting a shogun exception again again when i apply the trained classifier to my test set.05:44
ishaanmlhtrany suggestions on how to go about it?05:45
Ram108no sorry i havent reached there yet....... i have just finished the train phase...... am gona work on that one now05:46
ishaanmlhtrok..tell me if you are done with it.also,any idea why the parameter_list got 2 sets of arguments?05:48
ishaanmlhtri am talking about the kNN classifier file.05:48
ishaanmlhtrin python_modular examples05:48
ishaanmlhtrhave you gone through it?05:48
Ram108parameter_list = [[traindat,testdat,label_traindat,3].....?05:49
ishaanmlhtrif i remove the 2nd argument,then also it works,no?05:50
Ram108apparantly both are the same....... no idea...... i ll try figuring that out tweaking parameters here and there05:51
-!- ishaanmlhtr [~ishaan@] has quit [Quit: Leaving]06:11
-!- blackburn [~blackburn@] has joined #shogun10:21
-!- Ram108 [~amma@] has quit [Quit: Ex-Chat]11:22
-!- Ram108 [~amma@] has joined #shogun12:48
-!- ishaanmlhtr [~ishaan@] has joined #shogun13:50
blackburnishaanmlhtr: I sent ;)13:52
ishaanmlhtrblackburn : ya, got it..thanks :)13:53
Ram108hi blackburn13:53
blackburnRam108: hi13:53
Ram108sorry had exams last weekcouldnt catch up13:54
Ram108i ll start on it soon :)13:54
ishaanmlhtrblackburn : could you tell me some simplest application that i could try implementing along with implementing the other classifiers?13:54
blackburnRam108: no problem, feel free to do anything :)13:55
blackburnishaanmlhtr: ehm do you want to implement some classifier?13:55
blackburnwhich one?13:56
ishaanmlhtrgaussian bayes?13:57
blackburnalready here13:57
blackburnif yes then already here13:57
blackburnishaanmlhtr: I can suggest you some application with python13:57
blackburnishaanmlhtr: there is a MNIST dataset:
ishaanmlhtri should go through it first?13:58
blackburnishaanmlhtr: it would be nice to have an application which runs on this dataset13:59
blackburnand compares different methods13:59
blackburnif you want to implement some classifier or so I can suggest you something too13:59
blackburnit is up to you13:59
ishaanmlhtri think i'll have a read first. in any case, i am trying out the different classifiers that are already there in shogun on the flowers dataset14:00
blackburnyeah good practice14:00
blackburnbut iris dataset is small14:00
blackburnwhile MNIST is huuuuuge14:01
ishaanmlhtrok,then that's better - i can try the classifiers now on MNIST14:01
ishaanmlhtrand compare the results even more accurately.14:01
ishaanmlhtri'll have a read and get back to you in case i am stuck up.14:02
blackburnishaanmlhtr: I can send you a snipped I've used to read this data format14:02
ishaanmlhtrya,sure any help would be good.:)14:03
ishaanmlhtrplease mail me.thanks.:)14:03
blackburnishaanmlhtr: sent. you have to put data in mnist/ folder right with that script to run14:05
-!- puneetgoyal [~puneetgoy@] has joined #shogun14:30
-!- blackburn [~blackburn@] has quit [Quit: Leaving.]16:01
-!- blackburn [~blackburn@] has joined #shogun16:02
-!- ishaanmlhtr [~ishaan@] has quit [Ping timeout: 248 seconds]17:33
-!- ishaanmlhtr [~ishaan@] has joined #shogun17:52
-!- Ram108 [~amma@] has quit [Quit: Ex-Chat]18:06
blackburnsonney2k: did you manage to glance over my paper?19:31
-!- axitkhurana [79f5167c@gateway/web/freenode/ip.] has joined #shogun19:43
-!- blackburn [~blackburn@] has left #shogun []20:02
-!- ishaanmlhtr [~ishaan@] has quit [Quit: Leaving]20:26
-!- puneetgoyal [~puneetgoy@] has quit [Quit: Leaving]20:39
-!- blackburn [~blackburn@] has joined #shogun21:30
-!- axitkhurana [79f5167c@gateway/web/freenode/ip.] has quit [Quit: Page closed]21:56
--- Log closed Mon Dec 19 00:00:19 2011