| --- Log opened Sun Apr 13 00:00:03 2014 | ||
| -!- khalednasr [~k.nasr92@41.69.160.123] has quit [Quit: Leaving] | 00:10 | |
| -!- HeikoS [~heiko@90.201.62.238] has joined #shogun | 00:20 | |
| -!- mode/#shogun [+o HeikoS] by ChanServ | 00:20 | |
| -!- lisitsyn [~qdrgsm@109-226-124-25.clients.tlt.100megabit.ru] has quit [Quit: Leaving.] | 00:30 | |
| -!- lisitsyn [~qdrgsm@109-226-124-25.clients.tlt.100megabit.ru] has joined #shogun | 00:32 | |
| -!- diver-in-sky_ [uid28350@gateway/web/irccloud.com/x-mddxzutqvfaondva] has quit [Quit: Connection closed for inactivity] | 00:35 | |
| -!- lambday [67157d51@gateway/web/freenode/ip.103.21.125.81] has joined #shogun | 00:40 | |
| lambday | HeikoS: hi | 00:40 |
|---|---|---|
| lambday | there? | 00:40 |
| @HeikoS | lambday: hey | 00:40 |
| @HeikoS | yes | 00:40 |
| @HeikoS | how are things? | 00:40 |
| @HeikoS | in fact was just about to write you an email | 00:40 |
| lambday | HeikoS: a bit slow | 00:40 |
| lambday | :D | 00:40 |
| lambday | HeikoS: this research took an exciting turn! | 00:41 |
| lambday | HeikoS: will it make into jmlr? | 00:41 |
| @HeikoS | lambday: on the long run that might happen | 00:42 |
| @HeikoS | lambday: for now, we have some other things in mind. did you see the abstract? | 00:42 |
| lambday | HeikoS: yeah | 00:42 |
| lambday | I didn't know what to comment :( | 00:42 |
| @HeikoS | lambday: ah yeah read the email | 00:43 |
| @HeikoS | lambday: you dont have to for now | 00:43 |
| @HeikoS | yeah JMLR is the target goal | 00:43 |
| @HeikoS | which would be amazing right? | 00:43 |
| lambday | yeah man!! | 00:43 |
| lambday | didn't even think that! | 00:43 |
| @HeikoS | haha :) | 00:44 |
| lambday | so all the experiments should be redone and we'll use our shogun code | 00:44 |
| @HeikoS | yeah exactly | 00:44 |
| lambday | ramstein song ;) | 00:44 |
| @HeikoS | thats what my email will be about | 00:44 |
| @HeikoS | some things have to be changed | 00:44 |
| lambday | okay... I was confused a bit regarding how to fit things exactly... | 00:44 |
| lambday | will have to discuss after reading your mail | 00:45 |
| @HeikoS | what do you mean? | 00:45 |
| lisitsyn | night chat! | 00:45 |
| lisitsyn | :D | 00:45 |
| lambday | lisitsyn: nightbirds... :D | 00:45 |
| lambday | 4:15AM here | 00:45 |
| @HeikoS | lambday: for pre-gsoc, we should focus on getting this testing stuff done, but feature selection will be very similar in some ways | 00:45 |
| lisitsyn | % - 2hrs | 00:45 |
| @HeikoS | lambday: ah man, you dont sleep ;) | 00:45 |
| lambday | HeikoS: I wake up at 1 :D | 00:45 |
| @HeikoS | I see :) | 00:46 |
| lambday | HeikoS: alright... I am on it | 00:46 |
| @HeikoS | lambday: Ill just write mail. 1 sec | 00:46 |
| lambday | HeikoS: I was trying to draw a class diagram in this refactored scenario... not quite complete though | 00:46 |
| @HeikoS | or in fact, 10 mins if you have :) | 00:46 |
| lambday | yeah | 00:46 |
| lambday | yeah I am here | 00:46 |
| @HeikoS | lambday: ok sent | 01:04 |
| lambday | HeikoS: checking | 01:04 |
| lambday | HeikoS: alright... I get the points | 01:09 |
| lambday | there might be something that I am not currently seeing at this moment.. will discuss when I get there | 01:10 |
| lambday | and for streaming stuffs.. you meant that we should always rely on online mean/var algorithms, right? | 01:11 |
| lambday | so no storing but keep on computing | 01:11 |
| lambday | sum of kernel values without diag seems fine... two methods is better than flag | 01:12 |
| @HeikoS | yeah | 01:13 |
| @HeikoS | lambday: and then change quadratic time mmd thing | 01:13 |
| @HeikoS | to just call sum | 01:13 |
| @HeikoS | should be quite a bit faster when a custom kernel is used | 01:13 |
| @HeikoS | for example for the permutation test | 01:13 |
| lambday | hmm yeah | 01:14 |
| lambday | HeikoS: I am a bit confused about what to do with H_A variances... | 01:15 |
| lambday | I mean, for example, we would have methods that computes \sigma^2_{k, A} in eq (2) of Dino's writeup | 01:16 |
| lambday | right? | 01:16 |
| lambday | but when do we actually use that? | 01:17 |
| @HeikoS | for kernel selection | 01:17 |
| lambday | okay | 01:17 |
| @HeikoS | lambday: dont worry about that for now | 01:17 |
| lambday | alright | 01:17 |
| @HeikoS | for the testing, we just need the variance under the null | 01:17 |
| lambday | so p-value threshold they deal with var under null | 01:18 |
| lambday | yeah | 01:18 |
| @HeikoS | yes exactly | 01:18 |
| @HeikoS | so need to correct a few things, now, but keep the old ways in this deprecated methods | 01:18 |
| lambday | okay | 01:19 |
| lambday | and what do we check our new codes against? should we make a matlab implementation or so? | 01:19 |
| @HeikoS | lambday: I might have actually been wrong with one thing | 01:19 |
| lambday | HeikoS: what? | 01:19 |
| @HeikoS | the within block direct restimation contains a matrix product | 01:19 |
| lambday | yeah | 01:20 |
| @HeikoS | but maybe its possible to do that wihtout storing the matrix somehow, dont know, maybe think about it | 01:20 |
| lambday | ummm... but its just the kernel matrix for the current block, right? | 01:21 |
| lambday | I mean, how huge that would be? | 01:21 |
| lambday | and for the sum.. we need to specify blocks... because the kernel matrix is formed with merged samples of p_and_q | 01:25 |
| lambday | for one part we don't need diag elements.. for some parts we need | 01:25 |
| lambday | anyway I'll think about it | 01:26 |
| @HeikoS | lambday: think big | 01:31 |
| @HeikoS | lambday: blocks might be larger than 10000 which is maximum matrix size one can store | 01:32 |
| @HeikoS | lambday: also efficiency! if you dont have to store things, dont store them if its not slower | 01:32 |
| @HeikoS | and yes, feel free to adjust the interface of this sum operation to your needs | 01:33 |
| @HeikoS | lambday: maybe it should even be average operation, have to think about a clean interface for this and via not using too many if then else statements in the middle | 01:33 |
| @HeikoS | lambday: I will leave soon, let me know if you ahve more quetsions | 01:34 |
| @HeikoS | lambday: and yes, another python draft of things to uniut test against and for us to proofread (easier in python than in c++) might be very helpful | 01:34 |
| lambday | HeikoS: alright... | 01:36 |
| lambday | HeikoS: I'll think about this a bit.. will mail you regarding further queries | 01:36 |
| @HeikoS | yeah pls ask many questions! | 01:36 |
| lambday | for permutation test we got to sacrifice some power.. gotta store stuffs else it would be really slow | 01:38 |
| @HeikoS | lambday: what do you mean? | 01:38 |
| @HeikoS | lambday: ah yes, of course | 01:38 |
| @HeikoS | lambday: so the thing about the permutation test is: | 01:38 |
| @HeikoS | if possible: store, and dont call ::kernel in the loop, but rather let CCustomKernel loop over matrix | 01:39 |
| @HeikoS | so try to precompute matrix in any case | 01:39 |
| @HeikoS | like its now, it replaces the kernel by a custom kernel for the bootstrapping, you probably saw that | 01:39 |
| lambday | yeah | 01:39 |
| @HeikoS | lambday: but if its not possible to allocate the memory for that | 01:39 |
| @HeikoS | it should still be possible to do the test | 01:40 |
| @HeikoS | so have a boolean setter do de-activate that | 01:40 |
| lambday | alright | 01:40 |
| @HeikoS | and in fact, it would be cool if the precoimputing method would catch out of mem errors, print a warning that it will switch, and then do on-the-fly computing | 01:40 |
| @HeikoS | and then if one sets this precomputign flag to false, the warning never appears | 01:41 |
| lambday | yeah that would be cool! | 01:41 |
| @HeikoS | with that, the test should work for 100000 and more points (we should test that) | 01:41 |
| @HeikoS | you can do unit test with 3 bootstrap iterations for such cases ;) | 01:41 |
| lambday | :D | 01:41 |
| lambday | alright... so I am thinking that starting with quadratic time mmd and its related dependencies (like the kernel sum/mean computation etc) would be good | 01:43 |
| lambday | HeikoS: then we'd deal with streaming cases | 01:44 |
| lambday | HeikoS: btw with the refactored design with many methods pulled out from LinearTimeMMD to StreamingMMD, I think we'd lose those preallocated vector comprised methods... but if those can be done in parallel then that's a sacrifice worth of | 01:45 |
| @HeikoS | lambday: which methods do you mean? | 01:47 |
| @HeikoS | for streaming, no need to store anything right? | 01:47 |
| @HeikoS | just the block-size | 01:47 |
| lambday | HeikoS: yeah.. wait let me show | 01:47 |
| @HeikoS | where data has to be stored | 01:47 |
| @HeikoS | and kernel should probably not be stored (one test only computes statistics once, one does twice, one has this matrix product where we have to see) | 01:47 |
| lambday | HeikoS: yeah... but its not related to that | 01:48 |
| @HeikoS | lambday: multicore parallelisation might be interesting, but not for now btw | 01:48 |
| lambday | okay | 01:49 |
| @HeikoS | lambday: first goal is to make things stable and efficient | 01:49 |
| @HeikoS | then we can speed up | 01:49 |
| @HeikoS | lambday: keep in mind precomputing kernel matrices is already parallelised | 01:49 |
| lambday | yeah what I was saying was related to multiple kernel stuff... in the method compute_statistic_and_Q | 01:50 |
| @HeikoS | lambday: ah yeah | 01:50 |
| lambday | so it has these preallocated vectors... | 01:50 |
| @HeikoS | (that might be good to do in parallel in fact) | 01:50 |
| @HeikoS | ah thats fine | 01:50 |
| lambday | yeah | 01:50 |
| @HeikoS | not many kernels so storing multiple there is pok | 01:50 |
| @HeikoS | and yeah, this in fact SHOULD be done in parallel if easily possible | 01:51 |
| lambday | yeah... | 01:51 |
| @HeikoS | lambday: but not too important, getting things to work is the first step | 01:51 |
| @HeikoS | and have thorough unit tests | 01:51 |
| @HeikoS | so that parallelisation can be tested easily, which makes it easier to develop it | 01:52 |
| lambday | okay... so you agree with starting with quadratic time mmd ? | 01:52 |
| @HeikoS | yes! | 01:52 |
| lambday | although it may be useful to first think it all before changing - I'll do some planning tomorrow | 01:53 |
| @HeikoS | lambday: yeah | 01:53 |
| @HeikoS | lambday: planning is good | 01:53 |
| @HeikoS | but maybe just to the point where you know how to do things in a way that its not woo much work to change to multicore | 01:54 |
| @HeikoS | spending time with pthreads would be too much ;) | 01:54 |
| lambday | hehe | 01:54 |
| lambday | oh and for all these tests, we'll have two settings each, right? full and incomplete | 01:54 |
| @HeikoS | yep | 01:54 |
| lambday | so enum? | 01:54 |
| @HeikoS | the incomplete ones are not important for now though | 01:55 |
| @HeikoS | yeah enum | 01:55 |
| lambday | and additional variance_under_alternative() methods for all | 01:56 |
| @HeikoS | yeah, this can sometimes probably be computed on the fly with other thingfs | 01:58 |
| @HeikoS | see linear time mmd | 01:58 |
| lambday | yeah | 01:58 |
| lambday | I meant whether we need a wrapper like compute_variance() that internally calls compute_statistic_and_variance() | 01:59 |
| @HeikoS | lambday: I think its always a good idea to modularise | 02:05 |
| @HeikoS | lambday: ok going to sleep now, see you later! | 02:05 |
| lambday | HeikoS: okay | 02:05 |
| lambday | HeikoS: yeah me too! | 02:05 |
| lambday | see you :) | 02:06 |
| lambday | gn | 02:06 |
| @HeikoS | lambday: you should ;) | 02:06 |
| lambday | :D | 02:06 |
| @HeikoS | bye, good night! | 02:06 |
| lambday | bye | 02:06 |
| -!- lambday [67157d51@gateway/web/freenode/ip.103.21.125.81] has quit [] | 02:06 | |
| @HeikoS | lisitsyn: still there? | 02:15 |
| -!- HeikoS [~heiko@90.201.62.238] has quit [Ping timeout: 245 seconds] | 02:21 | |
| -!- lisitsyn [~qdrgsm@109-226-124-25.clients.tlt.100megabit.ru] has quit [Ping timeout: 276 seconds] | 02:46 | |
| -!- lisitsyn [~qdrgsm@109-226-124-25.clients.tlt.100megabit.ru] has joined #shogun | 03:03 | |
| shogun-buildbot | build #673 of nightly_all is complete: Failure [failed compile test] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_all/builds/673 | 03:18 |
| -!- khalednasr [~k.nasr92@41.69.137.37] has joined #shogun | 04:06 | |
| -!- khalednasr [~k.nasr92@41.69.137.37] has quit [Client Quit] | 04:11 | |
| -!- kislay|exams [~Abhijeet@117.239.94.112] has quit [Read error: Connection reset by peer] | 05:04 | |
| -!- Netsplit *.net <-> *.split quits: lisitsyn, sonne|work, pfm_, shogun-buildbot, @sonney2k, vvu|Log, soumyaC, @wiking | 05:33 | |
| -!- Netsplit *.net <-> *.split quits: chakri, [BNC]dhruv13j, @besser82, zxtx, @ChanServ, naywhayare | 05:33 | |
| -!- Netsplit over, joins: vvu|Log | 05:39 | |
| -!- sonney2k [~shogun@7nn.de] has joined #shogun | 07:16 | |
| -!- besser82 [quassel@gateway/shell/bnc4free/session] has joined #shogun | 07:16 | |
| -!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 07:16 | |
| -!- zxtx [~zxtx@sunnythinking.org] has joined #shogun | 07:16 | |
| -!- naywhayare [~ryan@spoon.lugatgt.org] has joined #shogun | 07:16 | |
| -!- ServerMode/#shogun [+o sonney2k] by sendak.freenode.net | 07:16 | |
| -!- lisitsyn [~qdrgsm@109-226-124-25.clients.tlt.100megabit.ru] has joined #shogun | 07:18 | |
| -!- [BNC]dhruv13j [dhruv13j@gateway/shell/bnc4free/x-zhvbyfoelcvhkhls] has joined #shogun | 07:18 | |
| -!- chakri [srichakrad@gateway/shell/bnc4free/x-cdbhudoxjppmjuli] has joined #shogun | 07:18 | |
| -!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 07:18 | |
| -!- ServerMode/#shogun [+o wiking] by sendak.freenode.net | 07:18 | |
| -!- sonne|work [~sonnenbu@91-64-78-194-dynip.superkabel.de] has joined #shogun | 07:18 | |
| -!- pfm_ [~hurr@hamilton.maths.tcd.ie] has joined #shogun | 07:18 | |
| -!- besser82 [quassel@gateway/shell/bnc4free/session] has quit [Changing host] | 07:19 | |
| -!- besser82 [quassel@gateway/shell/bnc4free/x-sjwrsczdlqfancsq] has joined #shogun | 07:19 | |
| -!- aditidatta [~aditidatt@117.194.118.32] has joined #shogun | 08:50 | |
| -!- aditidatta [~aditidatt@117.194.118.32] has quit [Client Quit] | 08:52 | |
| -!- ChanServ [ChanServ@services.] has joined #shogun | 09:18 | |
| -!- ServerMode/#shogun [+o ChanServ] by sendak.freenode.net | 09:18 | |
| -!- besser82 is now known as Guest41038 | 09:18 | |
| -!- Guest41038 is now known as besser82 | 09:54 | |
| -!- besser82 [quassel@gateway/shell/bnc4free/x-sjwrsczdlqfancsq] has quit [Quit: freedom, friends, features, first ---> fedoraproject.org] | 09:54 | |
| -!- besser82 [quassel@fedora/besser82] has joined #shogun | 09:56 | |
| -!- mode/#shogun [+o besser82] by ChanServ | 09:56 | |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 10:10 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 10:11 | |
| -!- sunil1337 [0e8bd057@gateway/web/freenode/ip.14.139.208.87] has joined #shogun | 11:07 | |
| -!- Netsplit *.net <-> *.split quits: @ChanServ | 11:18 | |
| -!- Netsplit over, joins: @ChanServ | 11:19 | |
| -!- PirosB3 [~pirosb3@host238-17-dynamic.245-95-r.retail.telecomitalia.it] has joined #shogun | 11:29 | |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 11:43 | |
| -!- sunil1337 [0e8bd057@gateway/web/freenode/ip.14.139.208.87] has left #shogun [] | 12:03 | |
| -!- lambday [67157d51@gateway/web/freenode/ip.103.21.125.81] has joined #shogun | 14:02 | |
| -!- pl8787 [9fe22b60@gateway/web/freenode/ip.159.226.43.96] has joined #shogun | 15:21 | |
| -!- pl8787 [9fe22b60@gateway/web/freenode/ip.159.226.43.96] has quit [Quit: Page closed] | 15:27 | |
| -!- pl8787 [9fe22b60@gateway/web/freenode/ip.159.226.43.96] has joined #shogun | 15:38 | |
| -!- khalednasr [~k.nasr92@41.69.186.4] has joined #shogun | 16:22 | |
| @wiking | lisitsyn: ping | 16:36 |
| lisitsyn | wiking: pong | 16:36 |
| @wiking | lisitsyn: 1) what happend with kaggle IO? | 16:36 |
| lisitsyn | wiking: not ready yet | 16:37 |
| lisitsyn | but I added downloaders | 16:37 |
| lisitsyn | wiking: will work on that soon | 16:37 |
| @wiking | 2) https://scan.coverity.com | 16:38 |
| @wiking | first login with github | 16:38 |
| lisitsyn | hmm | 16:38 |
| @wiking | https://scan.coverity.com/projects/1763?tab=Overview | 16:39 |
| @wiking | do u see there something like request more access? | 16:39 |
| lisitsyn | wiking: no | 16:39 |
| lisitsyn | just not authorized | 16:40 |
| @wiking | shit | 16:40 |
| @wiking | ok i guess pathen it's because it hasn't been fully configured... | 16:41 |
| lisitsyn | wiking: did you add shogun already? | 16:41 |
| @wiking | yeah i'm working on that | 16:42 |
| @wiking | but configuring with our crazy travis config | 16:42 |
| @wiking | is not the easiest task | 16:42 |
| @wiking | lisitsyn: but check out in the meanwhile the whole coverity stuff | 16:42 |
| @wiking | looks like a tool that we would benefit a lot from | 16:43 |
| lisitsyn | wiking: yeah | 16:43 |
| @wiking | but then again we should see the result first... | 16:43 |
| lisitsyn | true | 16:43 |
| lisitsyn | I have no idea yet | 16:43 |
| @wiking | i'll try to first do then the analysis locally | 16:44 |
| @wiking | doesn't have to be done via travis necessarily | 16:44 |
| @wiking | maybe we could even integrate a new buildbot task | 16:45 |
| @wiking | that runs every week once on the develop branch | 16:45 |
| @wiking | as otherwise we would need to maintain a new branch | 16:46 |
| @wiking | lisitsyn: what do you reckon? i feel like maybe actually integrating this to buildbot as a weekly once thing would be better than to maintain a new branch | 16:47 |
| @wiking | lisitsyn: https://scan.coverity.com/faq#frequency | 16:47 |
| lisitsyn | wiking: yeah sounds better for me too | 16:48 |
| lisitsyn | well all the thing is now like an overkill | 16:48 |
| @wiking | what do you mean? | 16:49 |
| lisitsyn | I mean we don't really need to do static analysis for each build and so on | 16:49 |
| @wiking | yeah we can change that as well | 16:49 |
| @wiking | lisitsyn: to a twice a week thing | 16:49 |
| -!- PirosB3 [~pirosb3@host238-17-dynamic.245-95-r.retail.telecomitalia.it] has quit [Quit: PirosB3] | 17:14 | |
| -!- PirosB3 [~pirosb3@host238-17-dynamic.245-95-r.retail.telecomitalia.it] has joined #shogun | 17:20 | |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 17:51 | |
| -!- mode/#shogun [+o iglesiasg] by ChanServ | 17:52 | |
| -!- PirosB3 [~pirosb3@host238-17-dynamic.245-95-r.retail.telecomitalia.it] has quit [Quit: PirosB3] | 17:58 | |
| -!- lambday [67157d51@gateway/web/freenode/ip.103.21.125.81] has quit [Ping timeout: 240 seconds] | 18:34 | |
| -!- khalednasr [~k.nasr92@41.69.186.4] has quit [Quit: Leaving] | 19:30 | |
| -!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 20:15 | |
| @wiking | lisitsyn: https://scan.coverity.com/projects/1763 | 20:37 |
| @wiking | lisitsyn: do u see the request button now? | 20:37 |
| lisitsyn | You are not authorized to access this page. | 20:39 |
| lisitsyn | wiking: ^ | 20:39 |
| @wiking | damn wtf | 20:39 |
| -!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 21:51 | |
| shogun-notifier- | shogun: abinashpanda :develop * 6b8f833 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/6b8f833e5253c1c7be340fae5464575c8d020639 | 21:51 |
| shogun-notifier- | shogun: CUAIFile class added to parse and write UAI file format | 21:51 |
| shogun-notifier- | shogun: tklein23 :develop * 0ade087 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/0ade087af368e4d14772b8db4fd1248e1a3956c6 | 21:51 |
| shogun-notifier- | shogun: Merge pull request #2019 from abinashpanda/develop | 21:51 |
| shogun-notifier- | shogun: | 21:51 |
| shogun-notifier- | shogun: Created CUAIFile class (Issue #1913) to save and load file in UAI file format. | 21:51 |
| -!- thoralf [~thoralf@91-65-136-198-dynip.superkabel.de] has joined #shogun | 23:03 | |
| thoralf | Good morning. | 23:03 |
| lisitsyn | thoralf: yeah that's a nice winter day :D | 23:14 |
| thoralf | lisitsyn: Still cold in russia? ;) | 23:21 |
| lisitsyn | thoralf: nah I mean if it is morning it could be a winter too ;) | 23:21 |
| thoralf | I know. I just needed an appropriate answer to you. :) | 23:23 |
| thoralf | It worked. :) | 23:23 |
| lisitsyn | thoralf: apart from that it is around 0C | 23:24 |
| thoralf | Really? | 23:25 |
| thoralf | We had 20 degree. | 23:25 |
| thoralf | 16 degree according to the forecast. | 23:25 |
| thoralf | 8 degree tonight. | 23:25 |
| lisitsyn | thoralf: it is on its way from -30 to +30 | 23:26 |
| lisitsyn | gradually | 23:26 |
| lisitsyn | half way done | 23:26 |
| thoralf | That's really bad. | 23:26 |
| thoralf | Come to Berlin. ;) | 23:27 |
| lisitsyn | thoralf: why bad? it is pretty cool actually | 23:27 |
| thoralf | It's between 10-20 most of the time. | 23:27 |
| lisitsyn | sometimes it is cold | 23:27 |
| lisitsyn | sometimes it is hot :D | 23:27 |
| lisitsyn | what sucks is that roads are ruined every spring | 23:27 |
| thoralf | We can choose between cold and warm rain. ;) | 23:28 |
| lisitsyn | thoralf: and you speak like I don't want to come to berlin ;) | 23:28 |
| lisitsyn | can't see the opportunity ;) | 23:28 |
| thoralf | lol | 23:28 |
| thoralf | Did you see this one? ;) https://github.com/shogun-toolbox/shogun/issues/2133 | 23:46 |
| thoralf | lisitsyn: Every time I touch shogun this happens. | 23:53 |
| lisitsyn | thoralf: actually sparse vector is assumed to be sorted | 23:53 |
| thoralf | lisitsyn: Well, if it's assumed to be sorted, why computing dot product this way? | 23:59 |
| thoralf | https://github.com/shogun-toolbox/shogun/blob/develop/src/shogun/lib/SGSparseVector.cpp#L355 | 23:59 |
| thoralf | You can do it in O(n), but this one is in O(n^2). | 23:59 |
| lisitsyn | unreadable :D | 23:59 |
| --- Log closed Mon Apr 14 00:00:04 2014 | ||
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!