IRC logs of #shogun for Thursday, 2018-02-22

« Wednesday, 2018-02-21 Index Friday, 2018-02-23 »
--- Log opened Thu Feb 22 00:00:12 2018
-shogun-buildbot:#shogun- Build nightly_all #99 is complete: Failure [failed test (failure)] - http://buildbot.shogun-toolbox.org:8080/#builders/22/builds/9903:56
-!- HeikoS [~heiko@host86-129-231-92.range86-129.btcentralplus.com] has joined #shogun10:22
-!- mode/#shogun [+o HeikoS] by ChanServ10:22
@wikingyou can write i there10:55
@wikingfor the future HeikoS10:55
@HeikoSsure10:56
@wikingi mean pretty much everything we would like to support10:56
@wikingwith std::vector<10:56
@wikingvector of floats of vectors of matrices of whatevers10:56
@wikingvector of SGOs10:56
@wikingmeaning it should be basically templated that much10:57
@wikingthat anybody could put anything into it10:57
@wikingsee for example cereal10:57
@wikingthat has these typemaps for STLs10:57
@HeikoSyes for al those10:57
@HeikoSI guess we start with std::vector10:58
@HeikoSand then see what happens10:58
@wikingmaps is something as well that should be able to hold most of things10:58
@HeikoSyes definitely need maps10:58
@wikingHeikoS, but yeah vector is the most imporntatn10:58
@wikingas that could eliminate a lot of handcrafted objects10:58
@HeikoSyep10:58
@wikingi dont see really what you mean by limitation10:59
@wikingof supporting10:59
@HeikoSmmh10:59
@wikingsay in the domain of std::vector10:59
@HeikoSstd::vector<std::vector<std::vector<CSGObject*>>10:59
@HeikoSthe old parameter framework had these weird things10:59
@HeikoSlike lists of matrices10:59
@HeikoSI would just start from what we need11:00
@wikingand why and how would you limit this?11:00
@HeikoSnot saying I would11:00
@HeikoSjust saying maybe there will be problems if these containers are nested to heavily11:00
@wikingi'm not so sure if i can follow11:00
@wikingi mean if you dont want to11:00
@wikingallow11:00
@wikingthen you'll have to do some sorts of magic hack11:01
@wikingwith typetraits11:01
@wikingif you do this with templates11:01
@HeikoSok then let's allow  :)11:01
@HeikoSthe comment wasn't too precise I guess :D11:01
@wikingi mean i just see the 2 ways11:01
@wikingeither you do some type trait hacking11:01
@HeikoSbtw I have a call very soon11:01
@wikingor you implement each and every case11:01
@wikingby hand11:01
@HeikoSnono11:01
@HeikoSplease no implementation by hand11:02
@HeikoSthere is already too much of that11:02
@wikingor you implement it with a template11:02
@wikingand then that can be triggered any way you want during compile time11:02
@HeikoSyep that sounds good11:03
@HeikoSlisitsyn:11:04
lisitsynhey11:04
@HeikoSlisitsyn: you saw the email?11:04
@HeikoSwith the date fernando suggested?11:04
lisitsynHeikoS: I am fine with that weekend11:05
@HeikoSah damn I am not actually11:06
lisitsyn:)11:06
@HeikoSok have a call now, will get back after11:06
@wikingi can only do weekend of 7-8 April11:07
lisitsynHeikoS: you can consider me free all along in april11:07
@wikingif that's somewhere between the budapest-belgrade axis :)11:07
lisitsynwiking: ah russian border you mean?11:07
@wikinghahahahah11:07
@sukey[https://github.com/shogun-toolbox/shogun] New branch reserialization created11:50
@sukey[https://github.com/shogun-toolbox/shogun] New commit https://github.com/shogun-toolbox/shogun/commit/827d8f886ed891e2b5c50de18aea95c1bfe37932 by vigsterkr11:50
-!- HeikoS [~heiko@host86-129-231-92.range86-129.btcentralplus.com] has quit [Quit: Leaving.]11:57
-!- zxtx_ [~zxtx@sunnythinking.org] has joined #shogun11:57
-!- zxtx [~zxtx@sunnythinking.org] has quit [Ping timeout: 264 seconds]11:59
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Ping timeout: 264 seconds]11:59
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun11:59
@sukey[https://github.com/shogun-toolbox/shogun] New commit https://github.com/shogun-toolbox/shogun/commit/6ff810e94af84ae47e3a1bd06a4abdf12de49ea0 by vigsterkr12:12
-!- shogun-buildbot_ [~shogun-bu@7nn.de] has joined #shogun12:25
-!- zoq_ [~marcus_zo@urgs.org] has joined #shogun12:29
-!- wiking_ [~wiking@huwico/staff/wiking] has joined #shogun12:30
-!- mode/#shogun [+o wiking_] by ChanServ12:30
-!- wiking [~wiking@huwico/staff/wiking] has quit [Disconnected by services]12:32
-!- wiking_ is now known as wiking12:32
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Ping timeout: 264 seconds]12:34
-!- zoq [~marcus_zo@urgs.org] has quit [Ping timeout: 264 seconds]12:34
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has joined #shogun12:51
-!- mode/#shogun [+o HeikoS] by ChanServ12:51
@sukey[https://github.com/shogun-toolbox/shogun] New commit https://github.com/shogun-toolbox/shogun/commit/0c952d8abdc7054845c3400a269ada1678e9cf0c by vigsterkr12:57
-!- travis-ci [~travis-ci@ec2-54-162-178-155.compute-1.amazonaws.com] has joined #shogun13:31
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/34475958513:31
-!- travis-ci [~travis-ci@ec2-54-162-178-155.compute-1.amazonaws.com] has left #shogun []13:31
@wiking387: [DEBUG] Own parameter Serializable::value= different from provided Serializable::value=13:38
@wiking387: [DEBUG] Own parameter DynamicObjectArray::array= different from provided DynamicObjectArray::array=13:38
@wikingfor some reason the visitor is not writing the value13:39
@wikingmmm it's the problem of the visitor actually :(13:49
@wiking365: [DEBUG] DynamicObjectArray(array=,dim1_size=,dim2_size=,dim3_size=,free_array=,resize_granularity=,use_sg_malloc=)13:49
@wiking365: [DEBUG] DynamicObjectArray(array=,dim1_size=,dim2_size=,dim3_size=,free_array=,resize_granularity=,use_sg_malloc=)13:49
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has quit [Ping timeout: 260 seconds]13:58
-!- travis-ci [~travis-ci@ec2-54-162-178-155.compute-1.amazonaws.com] has joined #shogun14:00
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/34475958514:00
-!- travis-ci [~travis-ci@ec2-54-162-178-155.compute-1.amazonaws.com] has left #shogun []14:00
@wikingthis used to work no lisitsyn ?14:00
lisitsynwiking: yes14:01
@wikingpfff14:01
lisitsynto_string worked :)14:01
lisitsynidk14:01
@wikingnow it only gets the tag14:01
@wikingbut not the value14:01
lisitsynshould be easy to fix14:02
@wikingi mean i'm just gdbing14:02
@wikingand dont get why the value is not there14:03
@wiking;)14:03
-!- HeikoS [~heiko@eduroam-int-pat-8-87.ucl.ac.uk] has joined #shogun14:13
-!- mode/#shogun [+o HeikoS] by ChanServ14:13
@wikingi mean14:17
@wikingthis has become some real beauty14:17
@wiking17042 wiking    20   0 4505424 4.136g 593660 R 100.0 13.2   1:07.43 gdb14:17
@wikingwtf is happening with this library14:17
@HeikoSthats the unit tests right?14:19
@wikingHeikoS, noup14:21
@HeikoShaha amazing14:21
@wikingHeikoS, simple meta integration tester14:21
@HeikoSneed plugins I guess14:21
@wikinglisitsyn, the storage type is foobar :D14:22
lisitsynstorage type?14:22
@wikingso we go into void on(...) {}14:22
@wiking:)14:22
lisitsynah14:22
@wikingdunno when this happened14:22
@wikingshogun::any_detail::typed_pointer<shogun::ArrayReference<shogun::CSGObject*, int> > (ptr=0x555555773bc0) at /home/wiking/shogun/src/shogun/lib/any.h:29414:23
@wikingon visitor->on(typed_pointer<T>(storage))14:23
@wikingright14:23
@wikingi guess this got broken on ArrayReference14:23
@wikingok that makes sense14:50
@wikingwe haven't defined for that14:50
@wikingbut still the rest is as well goes into on(...)14:51
@wikingoh i think i know14:51
@wiking:DDD14:51
@wikingas i think sometime you made lisitsyn the function arg const14:51
lisitsynahh14:51
lisitsynok14:51
@wikingbut return static_cast<T const*>(ptr);14:51
@wiking:)14:51
@wikingbut not the visitors14:51
@wiking:P14:51
@wikingok i'll fix it now14:52
lisitsyngood catch14:52
@wikingi dont know what to do with14:52
@wikingArrayReference14:52
@wikingdo we wanna have a [...] visitor for that? :)14:52
@sukey[https://github.com/shogun-toolbox/shogun] New commit https://github.com/shogun-toolbox/shogun/commit/7902ec7ee42389437d94362cc8094405d6d067cf by vigsterkr15:01
@wikingcoded blindly15:02
@wikinglets see what happens15:02
@wiking387: [DEBUG] Comparing parameter DynamicObjectArray::array of type shogun::ArrayReference<shogun::CSGObject*, int>.15:12
@wiking387: [DEBUG] Comparing parameter Serializable::value of type double.15:12
@wiking387: [DEBUG] Comparing using fequals<float64_t>(lhs, rhs).15:12
@wiking387: [DEBUG] Own parameter Serializable::value=0.95 different from provided Serializable::value=0.92515:12
@wiking387: [DEBUG] Own parameter DynamicObjectArray::array= different from provided DynamicObjectArray::array=15:12
@wikingok works15:12
@wikinglisitsyn, now this is super interesting15:36
@wiking[DEBUG] Comparing parameter DynamicObjectArray::array of type shogun::ArrayReference<shogun::CSGObject*, int>.15:36
@wiking[DEBUG] Comparing parameter VectorSerializable::value of type shogun::SGVector<double>.15:36
@wiking[DEBUG] All parameters of VectorSerializable equal.15:36
@wiking[DEBUG] Comparing parameter Serializable::value of type float.15:36
@wiking[DEBUG] Own parameter Serializable::value=0.95 different from provided Serializable::value=0.9515:36
@wiking[DEBUG] Own parameter DynamicObjectArray::array= different from provided DynamicObjectArray::array=15:36
@wiking:)15:36
lisitsynwiking: it uses no fequals15:37
lisitsynas you can see from the log15:37
@wikingyeah but this is the same when i'm runnning the same code15:37
@wikingwith cpp15:37
@wiking:D15:37
@wiking:boom:15:37
@wiking:D15:37
@sukey[https://github.com/shogun-toolbox/shogun] vigsterkr pushed 2 commits:15:39
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/5adcfd3307ccde6c0f77e01ec78b5454870d29c315:39
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/76b579fc5a71ea38fe1be656589d6c92ffa5d7ea15:39
@wikinglisitsyn, wonderfuuuuuulll15:44
@wikingvalue float32 0.949999988079071 vs value float64 0.9515:44
@wiking:)15:44
lisitsynwiking: constness again?15:44
@wikingi donkt know why is this actually happening15:45
@wikingi mean this is some typemap magic15:45
@wikingthe first one is python15:45
@wikingthe second is cpp15:45
@wiking'same' code15:45
@wikingok i dont get this15:47
@wiking(<type 'float'>, 0.95)15:47
@wiking(<type 'float'>, 0.95)accuracy = evals.evaluate(labels_predict, labels_test)15:47
@wikingi mean accuracy = evals.evaluate(labels_predict, labels_test)15:47
@wikingevals.evaluate will return surely a double15:47
@wikingnot a float :)15:47
@wikingvirtual float64_t evaluate(CLabels* predicted, CLabels* ground_truth);15:49
@wiking:)15:49
@wikingok i mean 'float' in python is double prec15:49
@wikingafaik15:49
@wikingmmm15:50
@wikingi think there's has to be a problem with put as well15:51
@wikingseems SWIG's type dispatcher15:51
@wikingcatches this float as float32_t15:51
@wikingyep15:53
@wiking:(15:53
@wikingHeikoS, ping me when you are here15:54
@wikinglisitsyn, because of this https://pastebin.com/f4n0mM7s15:55
@wikingput("asdf", pythonFloat) will always trigger to be float32_t15:55
@wikingif there's no overflow15:56
@sukey[https://github.com/shogun-toolbox/shogun] vigsterkr pushed 3 commits:16:07
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/1b47dc2938b437b31ec7677d22b0464a6b78c87916:07
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/d91fecad6695fa461b5ed7cae938bd4ebf19bc6d16:07
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/ffad70472592383c4909a115cd797792ee659d8116:07
-!- HeikoS [~heiko@eduroam-int-pat-8-87.ucl.ac.uk] has quit [Ping timeout: 240 seconds]16:29
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has joined #shogun16:43
-!- mode/#shogun [+o HeikoS] by ChanServ16:43
@wikinghaha16:44
@wikingi dont even know how is it possible that we set random seed, but still have differnet results16:44
@wiking:D16:44
@wikingnon-threadsafe random FTW16:45
@HeikoSwhats going on?17:41
@wikingHeikoS, so there's a problem with float32_t vs float64_t in python17:44
@HeikoSshogun or swig?17:44
@wikingif u have polymorphism17:44
@HeikoSwhich branch?17:44
@wikingin your swig/shogun interface17:44
@wikingnono in general17:44
@HeikoStypedo?17:44
@wikingthis is the case17:44
@wikingif it's a double that can fit into a float17:45
@wikingswig will auto convert it to a float17:45
@HeikoSah17:45
@HeikoSso if there is both methods17:45
@HeikoSit will always convert17:45
@wikingso if you have17:45
@HeikoSeven if one passes a double17:45
@wikinga(float32_t)17:45
@wikinga(float64_t)17:45
@wikingthen in the swig generated type dispatch17:45
@wikingif the passed func args fits into a float32_t17:46
@wikingit'll use that17:46
@HeikoSi see17:46
@HeikoSso is it a swig bug?17:46
@HeikoSor feature?17:46
@wikingfeature17:46
@wikinghttps://pastebin.com/f4n0mM7s17:46
@wikingthat's what swig generates17:47
@HeikoSI see17:47
@HeikoSkind of makes sense actually17:47
@HeikoShow can you know that your type is not wide enough, you check for overflow17:48
@wiking?17:48
@wikingi mean in python float = float64_t17:48
@wikingso i dont get why it needs to do this17:48
@wikingwhy it just doesn't used double17:48
@wikingif you want float32_t17:49
@wikingthen you need np.float32_t17:49
-!- travis-ci [~travis-ci@ec2-54-159-25-82.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/34483118217:49
-!- travis-ci [~travis-ci@ec2-54-159-25-82.compute-1.amazonaws.com] has left #shogun []17:49
@wikingas python doesnt have single prec float in it by default17:49
@HeikoSdoes it block you anywhere?17:50
@wiking?17:51
@wikingthis is a general problem17:51
@wikingfor polymorphism17:52
@wikingin target langs17:52
@HeikoSsure17:52
@wikinghttp://blog.dlib.net/2017/12/a-global-optimization-algorithm-worth.html17:56
@HeikoSyeah something like this would be coo lto have17:57
@HeikoSi wrote smth in these lines for some gsoc project17:58
@HeikoSthink I read this a while ago17:58
@wikingit's from december :)18:02
@HeikoSyeah for application I meant18:04
@HeikoS2018 :)18:04
@HeikoSIm calling it a day!18:04
@HeikoSsee you tomorrow18:04
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has quit [Ping timeout: 264 seconds]18:09
-!- shenghac [uid281668@gateway/web/irccloud.com/x-nxgvoqoyogpzaksq] has joined #shogun19:24
-!- zoq_ is now known as zoq19:29
@wikinglisitsyn, here?20:03
lisitsynwiking: hey20:09
lisitsynyes20:09
@sukey[https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4166 synchronized by durovo20:11
@wikingmmm i really dont understand how the fuck can the eigenvalues of 2 different solvers can be so far waya20:19
@wiking*away20:19
@wiking705: vector=[0.820716744468893866,2.48879797496342636]20:19
@wiking705: vector=[-1,1]20:19
lisitsynwiking: uh :D20:52
lisitsynwiking: -1, 1 sounds bad20:52
@wiking:D20:53
@wikingtell me about it20:53
@wiking:P20:53
@sukey[https://github.com/shogun-toolbox/shogun] vigsterkr pushed 4 commits:21:14
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/4af4ec29e8d2d17388b646a12291aa6bf832304221:14
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/20a4fa94b590db6b6e8c60744500b31f1450bf5d21:14
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/ba704455921725a8790ede35ee7f807764b0751c21:14
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/7db4f12205646e78b25ff7d3ff2656f2fd4cd8e021:14
-!- witness [uid10044@gateway/web/irccloud.com/x-eeknbaklfwlebfqp] has joined #shogun21:15
@sukey[https://github.com/shogun-toolbox/shogun] New commit https://github.com/shogun-toolbox/shogun/commit/d7583513753715549a9df6ef05cccaec71389210 by vigsterkr21:35
@sukey[https://github.com/shogun-toolbox/shogun] vigsterkr pushed 3 commits:21:43
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/01bce4b655def00a8baf401f30168ded7531846821:43
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/064cb995f16c0a21114e60fc8cff9af2bda7bece21:43
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/872ad4c63d4b78ce07fccc5268017a65361c199a21:43
@sukey[https://github.com/shogun-toolbox/shogun] New commit https://github.com/shogun-toolbox/shogun/commit/eb3ee42ba517d0086a1bb3cc18d7dfbb3b77fd57 by vigsterkr22:23
-!- travis-ci [~travis-ci@ec2-54-159-25-82.compute-1.amazonaws.com] has joined #shogun22:34
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/34497519622:34
-!- travis-ci [~travis-ci@ec2-54-159-25-82.compute-1.amazonaws.com] has left #shogun []22:34
-!- travis-ci [~travis-ci@ec2-54-159-25-82.compute-1.amazonaws.com] has joined #shogun22:50
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/34497519622:50
-!- travis-ci [~travis-ci@ec2-54-159-25-82.compute-1.amazonaws.com] has left #shogun []22:50
-!- travis-ci [~travis-ci@ec2-54-159-25-82.compute-1.amazonaws.com] has joined #shogun23:13
travis-ciit's Heiko Strathmann'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/34499237623:13
-!- travis-ci [~travis-ci@ec2-54-159-25-82.compute-1.amazonaws.com] has left #shogun []23:13
-!- HeikoS [~heiko@host86-129-231-92.range86-129.btcentralplus.com] has joined #shogun23:22
-!- mode/#shogun [+o HeikoS] by ChanServ23:22
-!- HeikoS [~heiko@host86-129-231-92.range86-129.btcentralplus.com] has quit [Quit: Leaving.]23:31
-!- travis-ci [~travis-ci@ec2-54-161-227-244.compute-1.amazonaws.com] has joined #shogun23:36
travis-ciit's Heiko Strathmann'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/34499237623:36
-!- travis-ci [~travis-ci@ec2-54-161-227-244.compute-1.amazonaws.com] has left #shogun []23:36
@sukey[https://github.com/shogun-toolbox/shogun] vigsterkr pushed 15 commits:23:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/4ec68aa3e6d4dd17d610f0a256eeecf8ae3c271123:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/3b9f989d4938317d5bd127daa382c75447eb923e23:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/5460d86050f418804961a7c379f558859cdfd10923:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/1d15fc9a105c02cfed615fd27e44b58a320a486e23:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/fa9929681e3a2efbc52a31e64115672f946e8ee823:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/bbfa8013475c957e6896170dbf8f04f9eea80d0f23:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/d5a5d2b66f27ba47edeaead5220fd12a10b8bb1223:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/9da37a260e15eb2ba112a865b6290b2635bdad9223:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/eacc28a15d6b20c4258c32fa7f0dab5f3d5944a623:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/8524a78356481bd502c10df25137d0962e6f5f7623:47
@sukeyhttps://github.com/shogun-toolbox/shogun/commit/2270774f110d9bdbadd53be97ac5ee09ea405c8323:47
-!- sukey [~nodebot@ks312251.kimsufi.com] has quit [Remote host closed the connection]23:47
--- Log closed Fri Feb 23 00:00:14 2018
« Wednesday, 2018-02-21 Index Friday, 2018-02-23 »

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