IRC logs of #shogun for Wednesday, 2017-08-09

« Tuesday, 2017-08-08 Index Thursday, 2017-08-10 »
--- Log opened Wed Aug 09 00:00:31 2017
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: ZNC 1.6.3 - http://znc.in]03:44
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun04:50
-!- mode/#shogun [+o wiking] by ChanServ04:50
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3954 synchronized by olinguyen05:41
@wikingolinguyen, oh you are still awake? :)05:42
olinguyenyea lo05:42
olinguyenlol*05:42
@wikingthnx for the change05:42
olinguyennp!05:43
@wikingas soon as it's green i'll merge it but it should be green :)05:43
olinguyenawesome, thanks!05:43
@wikingnw05:44
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3960 synchronized by MikeLing06:03
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3954 synchronized by olinguyen07:29
-!- olinguyen [81615ad9@gateway/web/freenode/ip.129.97.90.217] has quit [Quit: Page closed]07:36
-!- mikeling [uid89706@gateway/web/irccloud.com/x-vsniquqegldoincf] has joined #shogun09:09
mikelingwiking: ping09:09
@wikingping09:09
@wikingpong09:09
mikelingwiking: hi, does all these changes in kernel.h https://github.com/shogun-toolbox/shogun/pull/3960/files/520198d7b89d133b6602d9a39ff30666eb730347#diff-512b8f6e61391930b28fe03be5f01bdbR184 make sense to you?09:10
mikelingI just want to make sure these are necessary changes so I can change other places related to kernel. :)09:10
@wikinglemme check09:10
@wikingyes09:11
@wikingthose are necessary09:11
mikelingok great09:12
mikelingthank you09:12
-!- Netsplit *.net <-> *.split quits: sheogorath2710:49
mikelingwiking: what's this part of parameters for https://github.com/shogun-toolbox/shogun/blob/develop/src/shogun/distributions/HMM.h#L1331? Do we want them as index_t?11:19
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: ZNC 1.6.3 - http://znc.in]11:20
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun11:27
-!- mode/#shogun [+o wiking] by ChanServ11:27
@wikingmikeling, ok so that's a good question11:27
-!- sukey [~nodebot@ks312251.kimsufi.com] has quit [Remote host closed the connection]11:33
-!- sukey [~nodebot@ks312251.kimsufi.com] has joined #shogun11:34
-!- mode/#shogun [+o sukey] by ChanServ11:34
-!- sheogorath27 [sheogorath@gateway/shell/firrre/x-rdgcmdvvjpibutry] has joined #shogun11:34
-!- HeikoS [~heiko@host-92-0-169-11.as43234.net] has joined #shogun12:22
-!- mode/#shogun [+o HeikoS] by ChanServ12:22
@wikingmikeling, if you are here13:25
@wikingi'd say keep it int32t13:25
@wikingfor the time being13:25
@wikingHeikoS, man that hmm needs cleanup13:26
@wikingHeikoS, are you around by any chance?13:26
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3906 merged by vigsterkr13:28
@sukey[https://github.com/shogun-toolbox/shogun] vigsterkr pushed 10 commits:13:28
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/2d3324e5dd25e172f83f5b7712992555914b05b813:28
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/fc97d5d2f4dc09fdaac7b30be91e96022a83a35d13:28
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/e5a20fe8e67a54a465a1502cd561237791caff0c13:28
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/a3ebf996e016fb85c1f4356213d3b21a24d5523c13:28
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/80f7376150eb80981d9c47d1eded32edd3c28fdf13:28
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/b2276776cb52962ae62645d7ab36de42026435a413:28
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/c70f43d4304d0d65e035e042bf867ceac463b8d113:28
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/817eeeb3a40e3358b6cf7b52c44ca7fedf8311c213:28
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/5af0493da417d0d8e0fcfcb13f0a70dbcd960b5413:28
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/ef16d24f81fed90057991decce65c82c7e45983c13:28
@wikingmooooo13:28
@wikingaaaaa13:29
@HeikoSwiking: yes13:33
@wikingok so do you maybe have 10 mins?13:33
@HeikoSsure13:33
@HeikoSjust finished chapter 2 :D13:34
@wikinghahaha13:34
@wikingmaaan13:34
@wikingdont even remind me13:34
@wikinganyhow so i was just thinking about13:36
@wikingour solvers13:36
@wikinguff btw in the meanwhile i realised13:37
@wikingthat maybe we should restructure our unti test13:37
@wikingbecause now it's a bit stupid13:37
@wikingthat gpl related unit tests are in the main repo13:37
@wikingbut anyhow13:37
@HeikoSwiking: why?13:38
@wikingbecause it'd be better13:38
@HeikoSI thikn thats fine13:38
@wikingthat they are with13:38
@wikinggpl13:38
@HeikoSwhy?13:38
@HeikoSthe tests are gpl13:38
@HeikoSehm13:38
@wikingbecause they belong to each other?13:38
@HeikoSnot gpl13:38
@wikingi mean no use to have a source around13:38
@HeikoSbut the gpl repo is not even a real project13:38
@wikingif the corresponding13:38
@HeikoSjust source files13:38
@wikingyeah13:38
@HeikoSno cmake or anything13:38
@wikingbut we could foollow13:38
@HeikoSno build13:38
@wikingwhat most of the c++ project do nowadays13:38
@wikingespecially the bigger ones13:38
@HeikoSI really see this as a workaround for the gpl invasive nature13:38
@HeikoSso I see it as the same thing13:39
@HeikoSbut anyhow13:39
@HeikoSfeel free to change it13:39
@HeikoSthis is a minimal solution towars BSD so we can move on13:39
@wikingyeah i mean i was thinking about changing13:39
@wikingit as well13:39
-!- travis-ci [~travis-ci@ec2-54-211-49-115.compute-1.amazonaws.com] has joined #shogun13:39
travis-ciit's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/26262345013:39
-!- travis-ci [~travis-ci@ec2-54-211-49-115.compute-1.amazonaws.com] has left #shogun []13:39
@wikingbecause it's a bit shitty now13:39
@wikingtaht you have to move back and forth13:39
@wikingbetween13:39
@wikingdir structures13:39
@wiking(trees)13:39
@HeikoSit is13:39
@wikingfor the implementation and it's corresponding13:39
@wikinguntittest13:39
@wiking*unittest13:40
@wikingbigger c++ project13:40
@wikinginterleave these things13:40
@wiking_impl.cc _unittest.cc13:40
@HeikoSyeah sure13:40
@wikingif we'd have this13:40
@wikingthen we'd simply just put all the gpl stuff13:40
@wikinginto the gpl place13:40
@wikingand that's all13:40
@HeikoSno guarding13:40
@wikingyeah etc13:40
@HeikoSwell in practice, this is hard13:40
@wiking?13:40
@HeikoSbecause of how things are convoluted13:41
@wikingwhy hard?13:41
@HeikoSwith meta examples as well13:41
@wikingmmm yeah meta is a different story13:41
@wikingdont wanna touch that13:41
@wikingjust unittest for the time being13:41
@HeikoSI think it is ok as things are now actually13:41
@wikingk13:41
@HeikoSbut anyhow13:41
@wikingi'll change if you dont have any strong feelings13:41
@HeikoSwhat about the solvers?13:41
@wikingas i'd like to have13:41
@HeikoSno go ahead13:41
@wikingthe impl + unittest13:41
@wikingin one place13:41
@HeikoSsure do it if you want13:41
@wikingthis way we know as well13:41
@HeikoSI prefer that in fact13:41
@HeikoSbut I am also fine with it is now13:42
@wikingwha tthe hell is being tested13:42
@wikingkk13:42
@wikingso this just came up because i was looking for our friend13:42
@wikingDualLibQPBMSOSVM13:42
@wikingand the whole solver story in shogun13:43
@wikingand as well the solvers in tf for example13:43
@HeikoSyeah should be splot13:44
@HeikoSwith sep api13:44
@HeikoSit is so annoying that I cannot use a QP solver for the MMD stuff13:44
@HeikoSsince the only one that is there is GPL13:44
@HeikoSand the others are tangled with some other stuff I dont care about13:44
@wikingand then we have this beauty13:45
@wikinghttps://github.com/shogun-toolbox/shogun/blob/270ff9200daeb273bec6c82543d82d5a275b39d3/src/shogun/machine/StructuredOutputMachine.h#L12213:45
@HeikoSyeah13:45
@HeikoSI mean13:45
@HeikoSSO13:45
@HeikoSwtf13:45
@HeikoSthat is dead effectively13:45
@HeikoSthe community moved on to deep nets ;)13:45
@wikingmmm13:46
@wikingi dont see why it has to be dead13:46
@wiking:)13:46
@wikingi mean yeah it's note being used13:46
@wikingbut precisely for the fact13:46
@wikingthat for example there13:46
@HeikoSI remember how thoralf told me numerous times that this whole framework is fitting things into SVM framework with hammers13:46
@HeikoShaha13:46
@HeikoSanyhow13:46
@HeikoSwe have it13:46
@wikingyou had to generate your subgrad13:46
@wikingwhich is why everybody is13:46
@wikingfuck that13:46
@wiking:)13:46
@wikingthey are like man i have to code for my specific case the subgrad13:47
@wikinglalala13:47
@wikingbut i mean13:47
@wikingcouldn't we actually use now some nice solvers :)13:47
@HeikoSwould be cool to director class that13:47
@HeikoSsure13:47
@HeikoSthat would be cool13:47
@wikingthat would actually calculate13:47
@wikingthe subgrads13:47
@wikingbecause apparently we. could13:47
@wikingno?13:47
@wikingor am i being delusional?13:48
@HeikoShow?13:51
@HeikoSI don't know that framework too well13:51
@HeikoSbut aren't these things problem specific?13:51
@wikingwell yeah you need to define your13:52
@wikingdelta over the SO13:52
@wikingwhich is actually what you wanna most tweak13:52
@wikingand that is problem specific13:52
@HeikoSso what do you want to do there?13:53
@wikingand then there's the fact of how to represent the PSI(x,y)13:54
@wikingbut then it's a PO without any conditions13:54
@wikingso surely tf could solve that for us no?13:54
@wikingsorry13:55
@HeikoSyou mena the tf solvers?13:55
@wikingi'm in the meanwhile cooking and rebasing feature/random-refactor13:55
@HeikoSso you want to re-build the SO framework in tf effectively, which will give all the solvers from there, as well as the gradient stuff?13:55
@wikingyeah i mean now comes the question13:55
@wikinghow this would/should look like13:55
@wikingbecause surely we would like to use13:56
@wikingthe solvers one way or another13:56
@wikingfrom tf13:56
@wikingnot only in SO13:56
@wikingthe other somebody had a problem that was very SO-like13:56
@wikingremembered shit13:56
@wikingthen i realised that fuck with all those manual subgradient calculation/coding13:56
@wikingthis is horrible13:56
@HeikoSyep13:57
@HeikoSI get it13:57
@HeikoSwould be very nice13:57
@wikingbut then i was wondering if we could actually use TF here13:57
@wikingor stan for that matter13:57
@wikingwhichever would be the better fit13:57
@wikingi mean i have no idea what is exposed13:57
@wikingby stan's c++ api13:57
@wikingbut i guess13:57
@wikingyou can do anything you want no?13:58
@HeikoSyeah sure13:58
@HeikoSstan's api is not the nicest13:58
@HeikoSalso they dont have this graph stuff13:58
@HeikoSI dont think that is the way13:58
@HeikoStf is the way13:58
@wikingyeah tf was my immediate idea13:58
@HeikoSok then13:59
@HeikoSbut in order to do this13:59
@wikingyeah i mean we need optimizer framework13:59
@HeikoSwe need to have tf graphs in13:59
@HeikoSwhich is quite a task13:59
@HeikoSinterfacing with tf13:59
@HeikoSwould be SO useful for other things as well13:59
@HeikoSgps13:59
@wikingyeah13:59
@HeikoSI mean if you could have a linalg like thing with tf13:59
@HeikoSthat would do it13:59
@HeikoSbut we would need to copy the API behind a wrapper14:00
@HeikoSthen all this is possible14:00
@HeikoSor just directly use tf14:00
@HeikoSwithout wrapper14:00
@HeikoSbut I am hesitant with this as we had the problem of changing backends so many times now14:00
@wikingnono14:00
@wikingi would wrap it14:01
@HeikoSyeah ok14:01
@wikingbecause yeah14:01
@wikingwhat if there's tf++14:01
@HeikoSI mean that would be huge14:01
@wikingwhatevermachine14:01
@HeikoSgamechanger14:01
@wikingand we need to redo everything14:01
@wikingimo better just do a wrapping14:01
@wikingHeikoS, yeah i mean im just planning14:01
@HeikoSI would love to do that14:01
@wikingif things go well14:01
@HeikoSsince that really gives a perspective for shogun long term14:01
@wikingwhat could be a long running proj14:01
@HeikoSso yeah14:02
@HeikoSI discussed this a lot with others before14:02
@HeikoSbut I don't really know how to go ahead14:02
@wikingmmm14:02
@wikinggist .h14:02
@wiking:D14:02
@HeikoSwe could spend a hackathon prototyping stuff for that14:02
@wikingas usual14:02
@HeikoSI just wonder, is it the time for that already?14:02
@HeikoSor shouldnt we solve some other problems first14:02
@HeikoSlike the tags and cereal14:02
@wikingwhat would be more immediate?14:03
@wikingmmm tago14:03
@wikingyeah14:03
@wikingand being able to cereal14:03
@wikingyeah14:03
@wikingboth are valid14:03
@HeikoSand this stuff we discussed in zurich14:03
@wikingwhich?14:03
@HeikoSno templates in algorithms14:03
@wikingaah14:03
@HeikoSMatrix m = feats.get_cov14:03
@wikingbut yeah i wrote something about somewhere a bit14:03
@wikingregarding14:03
@HeikoSlinalg::invert(m)14:03
@wikingthe typedetection14:03
@HeikoSyeah14:03
@wikingand how not to do it again14:03
@wikingbecause the14:03
@wikingIF type if type if type14:03
@HeikoSyes that needs some clever stuff14:04
@wikingthat horrible14:04
@HeikoSI agree14:04
@HeikoSI mean but14:04
@HeikoSthe API doesnt care about that too much I think14:04
@HeikoSthis is backend stuff14:04
@wikingyeah14:04
@wikingsure i mean i just dont wanna do this behind the api14:04
@HeikoSso doing the API (e.g. this GSoC ) would be good14:04
@wikingthat in every function call14:04
@HeikoSyeah agreed14:04
@wikingthere's an if type switch14:04
@wiking:)14:04
@wikingkk14:04
@wikingbut having the api cleaned would be good14:04
@wikingactually14:04
@wikingam i right14:04
@wikingthat then those motherfucking14:04
@HeikoSmaybe you can push micmn a bit more on that14:04
@wikingmacros14:05
@wikingwould go?14:05
@HeikoSthe draft he made is useful14:05
@wikingbecause we dont need no more14:05
@HeikoScould put that in, change an algorithm or two to use it14:05
@HeikoSand then we can think about how to detect the types14:05
@wikingmacro it all the way the virtual functions14:05
@HeikoSI know, we can use annotators14:05
@HeikoS@float64_t14:05
@HeikoS:D14:05
@wikinglol14:05
@wiking:D14:05
@wikingwhish14:05
@wiking*wish14:06
@wikingthere's something coming14:06
@HeikoSbut basically we need to tell linalg: the next 10 operations are float6414:06
@wikingin c++20? :D14:06
@wikingyeah14:06
@HeikoSand it this is done via lazy eval, fine14:06
@HeikoSIll make lunch now14:06
@HeikoSwiking: but yeah can discuss this more14:06
@HeikoSI think micmn's patch is a good way to start this problem14:06
@HeikoStags where to start I dont know14:07
@HeikoSas we failed this GSoC :)14:07
@HeikoScereal depends on tags14:07
@HeikoSand then I am all game for TF integration14:07
-!- HeikoS [~heiko@host-92-0-169-11.as43234.net] has quit [Ping timeout: 248 seconds]14:13
Trixiswiking: shogun apache spark compatible version please:D14:48
@wikingTrixis, what is the incompatilibyt measure?14:48
Trixiswiking: nah i mean sparkify all the operations that can be sparkified :D14:48
@wikingokok14:48
@wikinggimme gimme ideas14:48
@wikingand then i can haz it14:49
@wikingbut yeah we dont have batch support for shit14:49
@wikingso you have to figure out yoruself14:49
@wikinga parameter server14:49
@wikingand batchify somehow your stuff14:49
@wikingright?14:49
Trixis(more just a dream for future really, ive been fuming that the cluster im using doesnt have spark)14:50
Trixiswiking: it seems to be that for example kernel learning could benefit from parallelisation across multiple machines via spark/hadoop/whatever14:50
Trixis*seems to me14:50
Trixissince some kernels / transforms are very expensive to compute, and the full matrix can require up to O(n^2) of such operations14:51
@wikingbut what14:51
@wikingyou would take the submatrix14:51
@wikingand do it batch by batch?14:52
@wikingand then reduce it ?14:52
Trixiswiking: um more of take the features, evaluate the individual kernel products, batch by batch, then reduce it into the matrix14:52
Trixiswiking: alternatively if youre training a large amount if individual learners, be it for evaluation or some combined classifier, you might as well do it batch by batch, then reduce it, no?14:54
@wiking:)15:03
@wikingyeah that's doable15:03
@wikingi guess only way is15:03
@wikingthat you have a view on the matrix15:03
@wikingright??15:03
@wikingi mean Features15:04
@wikingand then you do a batch15:04
@wikingmikeling, around?15:05
mikelingyes15:05
@wikingok so15:09
@wikingin15:09
@wikingjust a sec trying to find the commit15:09
mikelingok15:10
@wikinghttps://github.com/shogun-toolbox/shogun/pull/3906/commits/5af0493da417d0d8e0fcfcb13f0a70dbcd960b5415:11
-!- Saurabh7_ [Saurabh7@gateway/shell/panicbnc/x-qjqdjvhmwkfngaea] has quit [Ping timeout: 255 seconds]15:11
@wikinghere you removed CMath::permute15:11
@wikingbut didn't add it anywhere15:11
@wikingso currently as i've merged the pr into feature/random-refactor15:11
@wikingsrc/shogun/CMakeFiles/libshogun.dir/clustering/KMeansBase.cpp fails15:12
@wikingas it calls CMath::permute15:12
@wikinggit grep CMath::permute|wc -l15:12
@wiking      3815:12
@wikingso there are 38 places it's references15:12
@wiking*referenced15:12
-!- iglesiasg [~iglesiasg@217.119.234.214] has joined #shogun15:12
-!- mode/#shogun [+o iglesiasg] by ChanServ15:12
@wikingmikeling, what was there the idea? :)15:13
@wikingi mean now i'll force push that branch as i rebased it over the latest develop15:13
@wikingso you should send in a PR that fixes that15:13
@wikingok15:13
@wiking?15:13
mikelingmmmm, wait a second15:14
mikelingok, sure. I will send a pr for that15:14
@wikingk lemme test the rest15:15
mikelingbut actually , do i really remove the CMath::permute? I do remember we had discuss about that15:15
@wikingbecause the rebase wasn't an easy story15:15
@wikingmikeling, my idea is to move it to CRandom15:15
mikelingyou said maybe we should keep Random somehow for permute15:15
@wikingor that doesn't exist anymore?15:16
@wikingif not then we can move it to a utility function? :)15:16
@wikingi mean honestly15:17
@wikingmaybe it's just keep it there? :)15:17
@wikingmikeling, i can actually amend your last commit15:17
@wikingand just leave it there for the time being15:17
@wiking:P15:17
@wikingmikeling, any objections?15:17
mikelingno15:18
@wikingk15:18
@wikingi'll try to come up with the fixtures15:18
mikelingok, thank you :)15:18
Trixiswiking: hm i guess yeah15:21
@wikingok feature/random-refactor is rebased and should be good15:55
@sukey[https://github.com/shogun-toolbox/shogun] vigsterkr pushed 12 commits:15:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/306f0c6fe06349ecfff05725c49f58699094edb815:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/5d4c0372b7e33941db28bcb733cb6753fd07f5f615:55
@wikingnow it's only about the fixtures15:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/276a74d62ef785d6d3162b24dfeb7b8f45915a9e15:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/c81f0df50e9c4fd88122061f6ede83409fac245115:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/de234f6a778e36932d2a9860afc234bfe64abb0515:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/849740409ce85de7710fb234b2927fe6a6d510ef15:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/c0a7ef3c19117fc8fec60e9022907930208dee5515:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/02dfb1cb65238bdc60e2409aa74905cc0da55f9215:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/aa9cfc300d36bf89fcadd74e6e207f904183da4415:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/cc3eaaa340a7a9592fcc1bbd59adaf2109a3c55c15:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/35984d3be8faa336755ca24f362bc76067971b2d15:55
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/4f6831ce4739b772865b2a6a8afacf439af00b5215:55
@wikingsukey, flip15:55
@sukey‎(ﾉಥ益ಥ）ﾉ ┻━┻15:55
@sukey[https://github.com/shogun-toolbox/shogun] New commit https://github.com/shogun-toolbox/shogun/commit/58ef0574871acab3bbb322236036226fb27d7165 by vigsterkr16:08
-!- HeikoS [~heiko@host-92-0-169-11.as43234.net] has joined #shogun16:33
-!- mode/#shogun [+o HeikoS] by ChanServ16:33
-!- witness [uid10044@gateway/web/irccloud.com/x-qjwxxgenczokfsxk] has joined #shogun16:47
-!- travis-ci [~travis-ci@ec2-54-211-49-115.compute-1.amazonaws.com] has joined #shogun17:04
travis-ciit's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/26267492217:04
-!- travis-ci [~travis-ci@ec2-54-211-49-115.compute-1.amazonaws.com] has left #shogun []17:04
-!- olinguyen [81615ad9@gateway/web/freenode/ip.129.97.90.217] has joined #shogun17:29
olinguyenHeikoS: hey17:30
-!- travis-ci [~travis-ci@ec2-54-211-49-115.compute-1.amazonaws.com] has joined #shogun17:49
travis-ciit's Viktor Gal's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/26268016017:49
-!- travis-ci [~travis-ci@ec2-54-211-49-115.compute-1.amazonaws.com] has left #shogun []17:49
@HeikoSolinguyen: hi there17:49
olinguyenI can give you a rundown of the notebook here17:50
olinguyenI plotted the heart rate of 10 random patients that did not survive after 24 hours (shown in red), and 10 that did (shown in blue) to try and see if there's a significant difference.17:50
@HeikoSshall we have a chat?17:51
@HeikoSas in video?17:51
olinguyensure17:51
olinguyenlet me setup my laptop17:52
@HeikoSok17:52
@HeikoSolinguyen:  https://github.com/olinguyen/gsoc2017-shogun-dataproject/blob/master/Improved%20Model.ipynb17:56
@HeikoSolinguyen:  https://pandas.pydata.org/pandas-docs/stable/generated/pandas.Series.resample.html18:06
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/3954 synchronized by olinguyen18:32
-!- mikeling [uid89706@gateway/web/irccloud.com/x-vsniquqegldoincf] has quit [Quit: Connection closed for inactivity]18:38
@wikingok i guess this is now mergable then https://github.com/shogun-toolbox/shogun/pull/395419:22
@wikingright?19:23
@wikingif so i'll do a squash merge19:23
@HeikoSolinguyen: you still around?20:12
@HeikoSolinguyen: check your email20:13
@HeikoShttps://gist.github.com/karlnapf/f4ff9a69e5baede838212c2a9f5947ab20:13
@HeikoSlet me know if you have questions on this20:13
olinguyenok, going through it now20:13
@HeikoSif not, try to replicate this for multiple patients and try to train a model on it, testing it on different patients that you trained on20:13
olinguyenthanks for taking the time to do this20:13
@HeikoSolinguyen: I gotta leave now20:13
@HeikoSolinguyen: no worries20:13
-!- HeikoS [~heiko@host-92-0-169-11.as43234.net] has quit [Ping timeout: 255 seconds]20:19
-!- olinguyen [81615ad9@gateway/web/freenode/ip.129.97.90.217] has quit [Ping timeout: 260 seconds]21:28
shogitter(olinguyen) Cool22:26
-!- olinguyen [81615ad9@gateway/web/freenode/ip.129.97.90.217] has joined #shogun22:27
shogitter(olinguyen) test22:27
-!- witness [uid10044@gateway/web/irccloud.com/x-qjwxxgenczokfsxk] has quit [Quit: Connection closed for inactivity]22:37
-!- Saurabh7_ [Saurabh7@gateway/shell/panicbnc/x-xmqojrmdxdpwiivs] has joined #shogun23:51
--- Log closed Thu Aug 10 00:00:32 2017
« Tuesday, 2017-08-08 Index Thursday, 2017-08-10 »

Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!