--- Log opened Wed Nov 16 00:00:18 2016 | ||
-!- abhinav_ [~quassel@103.211.53.167] has joined #shogun | 02:43 | |
-!- abhinav_ [~quassel@103.211.53.167] has quit [Ping timeout: 268 seconds] | 05:44 | |
-!- abhinav_ [~quassel@103.211.53.167] has joined #shogun | 05:44 | |
-!- abhinav_ [~quassel@103.211.53.167] has quit [Ping timeout: 256 seconds] | 06:12 | |
-!- abhinav_ [~quassel@103.211.53.167] has joined #shogun | 07:04 | |
-!- sonne|work [~sonnenbu@88.134.176.135] has joined #shogun | 09:01 | |
-!- praisethemoon [~praisethe@41.230.52.174] has joined #shogun | 09:04 | |
-!- praisethemoon [~praisethe@41.230.52.174] has quit [Changing host] | 09:05 | |
-!- praisethemoon [~praisethe@unaffiliated/praisethemoon] has joined #shogun | 09:05 | |
-!- abhinav_ [~quassel@103.211.53.167] has quit [Read error: Connection reset by peer] | 09:13 | |
-!- abhinav_ [~quassel@103.211.53.167] has joined #shogun | 09:14 | |
lisitsyn | wiking: pluskid wrote a paper with hardt bengio et al | 09:21 |
---|---|---|
lisitsyn | impressive :) | 09:21 |
-!- praisethemoon [~praisethe@unaffiliated/praisethemoon] has quit [Ping timeout: 268 seconds] | 10:15 | |
-!- praisethemoon [~praisethe@41.230.52.174] has joined #shogun | 10:17 | |
-!- praisethemoon [~praisethe@41.230.52.174] has quit [Ping timeout: 268 seconds] | 10:24 | |
-!- praisethemoon_ [~praisethe@41.226.248.123] has joined #shogun | 10:24 | |
-!- praisethemoon_ is now known as praisethemoon | 10:25 | |
-!- praisethemoon [~praisethe@41.226.248.123] has quit [Changing host] | 10:25 | |
-!- praisethemoon [~praisethe@unaffiliated/praisethemoon] has joined #shogun | 10:25 | |
sonne|work | lisitsyn: did you expect anything less? | 10:47 |
lisitsyn | sonne|work: didn't like the paper tho | 10:47 |
lisitsyn | he attacks holy vapnik | 10:47 |
sonne|work | lisitsyn: do you have the url/pdf ? | 10:47 |
lisitsyn | sure | 10:47 |
lisitsyn | http://openreview.net/pdf?id=Sy8gdB9xx | 10:47 |
sonne|work | lisitsyn: attacking a god smells like immediate punishment ;) | 10:48 |
lisitsyn | sonne|work: you can't just attack the dark emperor | 10:48 |
lisitsyn | it's careless | 10:48 |
sonne|work | it is | 10:48 |
lisitsyn | sonne|work: tldr: neural nets overfit like crazy on random data and VC and Rademacher have no idea why it doesn't overfit when data is real | 10:49 |
lisitsyn | the section 5 is really strange | 10:50 |
lisitsyn | they reinvent KRR and say it works | 10:51 |
lisitsyn | UHMM OK :D | 10:51 |
praisethemoon | Hello everyone | 10:56 |
praisethemoon | I have been having a look at shogun lately, and seems very interesting :) | 10:56 |
praisethemoon | You guys are doing great job. | 10:56 |
-!- praisethemoon [~praisethe@unaffiliated/praisethemoon] has quit [Quit: Leaving] | 11:01 | |
-!- praisethemoon [~praisethe@41.230.52.174] has joined #shogun | 11:05 | |
-!- praisethemoon [~praisethe@41.230.52.174] has quit [Changing host] | 11:05 | |
-!- praisethemoon [~praisethe@unaffiliated/praisethemoon] has joined #shogun | 11:05 | |
sonne|work | lisitsyn: maybe because rademacher and VC only draw conclusions for all distributions not ones of natural property.. | 11:27 |
sonne|work | anyways even SVMs with gaussian kernels have VCdim == infty so... | 11:28 |
-!- abhinav_ [~quassel@103.211.53.167] has quit [Ping timeout: 250 seconds] | 11:32 | |
-!- praisethemoon_ [~praisethe@41.230.52.174] has joined #shogun | 12:09 | |
-!- praisethemoon [~praisethe@unaffiliated/praisethemoon] has quit [Ping timeout: 265 seconds] | 12:13 | |
-!- praisethemoon_ is now known as praisethemoon | 12:17 | |
-!- praisethemoon [~praisethe@41.230.52.174] has quit [Changing host] | 12:17 | |
-!- praisethemoon [~praisethe@unaffiliated/praisethemoon] has joined #shogun | 12:17 | |
-!- abhinav_ [~quassel@103.211.53.167] has joined #shogun | 13:13 | |
-!- Hieronymous [~sudarshan@122.169.172.64] has joined #shogun | 13:21 | |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has joined #shogun | 13:32 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 13:32 | |
@HeikoS | wiking: jo! | 13:35 |
-!- Hieronymous [~sudarshan@122.169.172.64] has quit [Remote host closed the connection] | 13:46 | |
-!- Hieronymous [~sudarshan@122.169.172.64] has joined #shogun | 13:48 | |
Hieronymous | HeikoS: jo! | 14:24 |
@HeikoS | jo | 14:25 |
Hieronymous | I am going through the PCA notebook. | 14:25 |
Hieronymous | But how to I create the cookbook ? I did read the docs, but its different for each type of learning problem. | 14:26 |
@HeikoS | I dont understand your question | 14:29 |
Hieronymous | HeikoS: Never mind, I got it. Any C++ beginner task available ? | 14:32 |
Hieronymous | Perhaps something in linalg ? I read the wiki, it is a bit "nascent". | 14:33 |
@HeikoS | Yeah you could remove some warnings | 14:36 |
@HeikoS | check our buildbot page | 14:36 |
@HeikoS | and the warnings for libshogun | 14:36 |
@HeikoS | need to be all resolved ;) | 14:37 |
@HeikoS | same for python | 14:37 |
@HeikoS | but I mean there is tons | 14:37 |
@HeikoS | check entrance tasks on our issues list | 14:37 |
@HeikoS | there is quite a few c++ ones | 14:37 |
@HeikoS | linalg needs more operations in the feature branch | 14:37 |
@HeikoS | check the commit history of the feature branch to get a feeling | 14:37 |
@HeikoS | eventially we want all operations in SGVector to be in linalg | 14:38 |
Hieronymous | Yeah, I am actually browsing through the entrance tasks. Nice I have something to research then. Great. | 14:38 |
-!- abhinav_ [~quassel@103.211.53.167] has quit [Remote host closed the connection] | 14:40 | |
@HeikoS | Hieronymous: yeah I mean there is so much stuff | 14:40 |
@HeikoS | just really pick anything that makes sense to you | 14:40 |
@HeikoS | the 5.1 milestone are the most pressing ones | 14:40 |
@HeikoS | there is some entrance ones in there as well | 14:40 |
Hieronymous | HeikoS: There are two feature branches for linalg , linalg_refactor and linalg_second_take. Which one should I be checking ? | 14:41 |
@HeikoS | refactor I think | 14:41 |
@HeikoS | the one OXPHOS has been active on | 14:41 |
@HeikoS | check the pull requests for that branch | 14:41 |
@HeikoS | they give you an idea of what is has gone through | 14:41 |
@HeikoS | but that is messy | 14:41 |
@HeikoS | rather wait a bit with that :) | 14:41 |
-!- lambday [6a331336@gateway/web/freenode/ip.106.51.19.54] has joined #shogun | 15:18 | |
-!- mode/#shogun [+o lambday] by ChanServ | 15:18 | |
@lambday | HeikoS: yo | 15:18 |
@HeikoS | lambday: jo man! | 15:18 |
@HeikoS | good to see you here :) | 15:18 |
@lambday | HeikoS: yeah :) | 15:18 |
@lambday | HeikoS: will spend some time on bigtesting now | 15:19 |
@HeikoS | cool | 15:19 |
-!- abhinav [~quassel@103.211.53.167] has joined #shogun | 15:19 | |
@lambday | HeikoS: BTW, the GPU kernels we have written may need revision | 15:19 |
@lambday | was looking at sum max, etc.. | 15:19 |
@lambday | the max work group size is hardcoded.. it should be queried from the h/w | 15:20 |
@lambday | anyway will do that later | 15:20 |
@lambday | maybe Pan has some ideas | 15:20 |
@lambday | for now, making bigtesting merge ready :) | 15:20 |
@HeikoS | yep absolutely | 15:20 |
@HeikoS | but have to do that *after* linalg is merged | 15:20 |
@HeikoS | same as for bigtest :) | 15:21 |
@HeikoS | currently, there is nothing | 15:21 |
@HeikoS | before we do fancy things, we want at least something :) | 15:21 |
@HeikoS | let me know how it goes and whether you need help | 15:21 |
@lambday | HeikoS: so let me know if I miss anything in the task list (1) make supplying a custom kernel easier, get rid of the extra info message (2) pre-initialize things with DummeFeatures.. the user is not expected to supply dummy featureds when he's providing a custom kernel (3) anything else? | 15:21 |
@lambday | will add the documentations next | 15:21 |
@HeikoS | yeah thats good | 15:22 |
@HeikoS | but (1) is: merge | 15:22 |
@HeikoS | before any of this is addressed | 15:22 |
@lambday | if some particular parameter is mandatory, it should be properly documented | 15:22 |
@HeikoS | yeah true | 15:22 |
@lambday | HeikoS: merge it today you mean? | 15:22 |
@lambday | HeikoS: if you have some time to spare we can do that today | 15:22 |
@HeikoS | no I mean merge before doing any of these points above | 15:22 |
@lambday | I'll rebase | 15:22 |
@HeikoS | I have | 15:22 |
@lambday | yeah I got that | 15:22 |
@HeikoS | the builds needs to pass | 15:23 |
@lambday | okay let me rebase | 15:23 |
@HeikoS | the notebook needs to work | 15:23 |
@lambday | and check the build | 15:23 |
@HeikoS | the hsic problem needs to be resolved | 15:23 |
@HeikoS | then docs | 15:23 |
@HeikoS | then merge | 15:23 |
@HeikoS | then your list above :) | 15:23 |
@lambday | added all to sticky notes | 15:25 |
@lambday | :D | 15:25 |
@HeikoS | nice one :) | 15:25 |
@HeikoS | we can tweet that ;) | 15:25 |
@lambday | HeikoS: shall I remove the HSIC files or just comment the whole thing out? | 15:25 |
@HeikoS | lambday: remove them | 15:27 |
@HeikoS | they are in the history | 15:27 |
@lambday | HeikoS: cool | 15:27 |
@sukey | Pull Request #3543 "Replace pointer and vlen with SGVector in QDA." opened by sudk1896 - https://github.com/shogun-toolbox/shogun/pull/3543 | 15:32 |
@lambday | HeikoS: the entire dir "statistics" should go, right? we have things in a different directory now | 15:32 |
@lambday | HeikoS: shall I merge this feature branch to my develop and then push changes? | 15:32 |
@lambday | I mean after I check things | 15:33 |
@HeikoS | nono | 15:33 |
@sukey | Pull Request #3543 "Replace pointer and vlen with SGVector in QDA." synchronized by sudk1896 - https://github.com/shogun-toolbox/shogun/pull/3543 | 15:33 |
@HeikoS | leave in feature branch in upstream | 15:33 |
@HeikoS | just rebase there | 15:33 |
@HeikoS | and polish there | 15:33 |
@HeikoS | I will merge from there | 15:33 |
@lambday | HeikoS: yeah but HSIC and those other things are not in that branch.. those have to be removed from develop itself | 15:34 |
@HeikoS | I dont understand | 15:34 |
@HeikoS | you have a branch that is rebased against develop | 15:34 |
@HeikoS | where you removed hsic | 15:34 |
@sukey | Pull Request #3543 "Replace pointer and vlen with SGVector in QDA." - https://github.com/shogun-toolbox/shogun/pull/3543 | 15:34 |
@HeikoS | if I merge this, hsic wil be deleted in develop | 15:34 |
@lambday | HeikoS: yeah sorry I got confused | 15:35 |
@lambday | HeikoS: so things are already deleted in my branch | 15:35 |
@lambday | BAHSIC and all are taken care of, since it does compile | 15:36 |
@HeikoS | remove all files/tests/folders/notebooks for that stuff as well | 15:36 |
@HeikoS | and then all good | 15:37 |
@lambday | HeikoS: hmm.. just tried rebasing now against develop.. got merge conflicts for two sample test notebook | 15:39 |
@lambday | fixing | 15:39 |
@HeikoS | yep that is since we had to clean things | 15:39 |
@HeikoS | before release | 15:39 |
@HeikoS | notebook was broken | 15:39 |
@lambday | I can replace it with the one in upstream in feature/bigtest branch, right? that one works with the new interface (will check whether it does) | 15:44 |
@lambday | nah it's not updated | 15:45 |
@lambday | changing | 15:45 |
@lambday | wow! github is broken | 15:47 |
-!- sonne|work [~sonnenbu@88.134.176.135] has quit [Ping timeout: 256 seconds] | 15:47 | |
-!- sonne|work [~sonnenbu@ip5886b087.fixed.kabel-deutschland.de] has joined #shogun | 15:48 | |
@HeikoS | happens every now and then :) | 15:50 |
@HeikoS | lambday: didnt get your question | 15:50 |
@lambday | HeikoS: gotta fix the notebook since I got merge conflicts.. there were so many.. was wondering whether I can replace the entire notebook from somewhere manually and then make changes according to our new API by running that thing | 15:53 |
@HeikoS | I think I updated the notebook quite a bit | 15:55 |
@HeikoS | in feature branch | 15:55 |
@HeikoS | to solve this is easy | 15:55 |
@HeikoS | take the feature branch version | 15:55 |
@HeikoS | and apply the latest changes from develop on top of it, they are in the history | 15:56 |
@HeikoS | I think these are only very few so should be easy | 15:56 |
@HeikoS | (i.e. change order of commits, dont rebase the notebook) | 15:56 |
@HeikoS | lambday: ^ | 15:56 |
@lambday | HeikoS: okay let me try | 15:57 |
@lambday | I took the feature branch version itself.. let me look at the latest commits in develop for that | 15:57 |
@lambday | there is just one that you did during the last release.. will apply that patch | 16:02 |
@HeikoS | exactly | 16:05 |
@HeikoS | and these are very minor changes I did for the release | 16:05 |
@lambday | HeikoS: umm.. it has 1,107 additions and 1,089 deletions.. if you fixed the develop one to make it work with the older code, then it might conflict with the one we have in our feature branch since the API changed | 16:13 |
@HeikoS | it is the format | 16:14 |
@HeikoS | you cannot apply the patch using git | 16:14 |
@HeikoS | have to do by hand | 16:14 |
@HeikoS | ah thats annoying | 16:14 |
@HeikoS | so I suggest | 16:14 |
@HeikoS | you check the changes I did for the release | 16:14 |
@lambday | yeah | 16:14 |
@HeikoS | which you cannot | 16:14 |
@HeikoS | since i did not make a sep patch for the format change | 16:15 |
@HeikoS | sigh | 16:15 |
@HeikoS | my mistake | 16:15 |
@HeikoS | so then, just discard my changes in develop | 16:15 |
@HeikoS | just make sure, that the notebook runs | 16:15 |
@HeikoS | (I did only that basically) | 16:15 |
@HeikoS | so just take your feature branch version, and period | 16:15 |
@lambday | HeikoS: cool.. just running the notebook | 16:17 |
@lambday | HeikoS: [- "version": "2.7.6"] [+ "version": "2.7.10"] | 16:19 |
@lambday | is this the change I should do? | 16:19 |
@lambday | umm | 16:19 |
@sukey | Pull Request #3544 "Replace pointers with SGVector" opened by souravsingh - https://github.com/shogun-toolbox/shogun/pull/3544 | 16:21 |
@HeikoS | lambday: no change | 16:24 |
@HeikoS | discard my stuff | 16:24 |
@HeikoS | it was only to make it executable again | 16:24 |
@HeikoS | just open the notebook in your own jupyter and run it | 16:24 |
@lambday | HeikoS: okay cool.. | 16:24 |
@lambday | yeah | 16:25 |
@HeikoS | (feeature branch version) | 16:25 |
@lambday | the develop version will anyway get removed since it's a different file in a new dir it seems | 16:25 |
@lambday | we removed it in feature branch | 16:25 |
-!- praisethemoon [~praisethe@unaffiliated/praisethemoon] has quit [Ping timeout: 268 seconds] | 16:28 | |
@HeikoS | lambday: what do you mean? | 16:32 |
@HeikoS | I thought you had conflicts | 16:32 |
@lambday | HeikoS: conflicts for the location where the feature branch version exists (!) - maybe because it was git mv'ed? | 16:33 |
@lambday | HeikoS: fixing this notebook now | 16:33 |
@sukey | Pull Request #3543 "Replace pointer and vlen with SGVector in QDA." - https://github.com/shogun-toolbox/shogun/pull/3543 | 16:33 |
Hieronymous | HeikoS: Take a look at the PR - > I cleaned up the QDA. Btw, tests passed. | 16:35 |
@HeikoS | Hieronymous: cool thanks, will check later | 16:35 |
-!- praisethemoon [~praisethe@41.226.248.123] has joined #shogun | 16:59 | |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has quit [Ping timeout: 258 seconds] | 17:05 | |
-!- praisethemoon [~praisethe@41.226.248.123] has quit [Quit: Leaving] | 17:58 | |
Hieronymous | lamday: Hey! | 18:10 |
Hieronymous | lambday, | 18:10 |
Hieronymous | Hi | 18:11 |
@lambday | Hieronymous: yo | 18:11 |
Hieronymous | Could you tell me how I could begin contrubuting to linalg ? | 18:11 |
@lambday | oh dear | 18:11 |
Hieronymous | I see that it is very "nascent". | 18:11 |
@lambday | umm.. we re-did all the internals and interfaces | 18:11 |
Hieronymous | Might be good for a noob such as myself. | 18:11 |
@lambday | TBH I gotta have a look at the current state myself, but Pan did an awesome job porting all those existing methods into the new framework | 18:12 |
@lambday | but still there are some left.. so that might be something you can look at | 18:12 |
@lambday | Hieronymous: do you have experience in OpenCL and GPU programming? | 18:13 |
Hieronymous | lamday: Nope. I was reading up a bit about Eigen though. | 18:13 |
Hieronymous | That seemed "intuitive" | 18:13 |
@lambday | Hieronymous: that's not a requirement, just curious cause we could have used your help :) | 18:13 |
Hieronymous | I am willing to learn new things. Shouldn't be a problem. | 18:14 |
@lambday | Hieronymous: no worries.. so you can collaborate with Pan moving the existing methods to the new interface.. | 18:14 |
Hieronymous | Where should I look for these "existing methods" | 18:14 |
@lambday | but maybe wait a bit, cause some of her patches are not yet merged.. you don't really want to work on something that somebody else did already | 18:15 |
Hieronymous | I did read the wiki on linalg and the reasons behind its inception. | 18:15 |
@lambday | Hieronymous: those are in the develop branch | 18:15 |
@lambday | check inside mathematics/linalg/internals | 18:15 |
@lambday | Hieronymous: Pan's changes are being merged into a feature branch linalg_refactor.. so that is another place to look at to know which methods are already there | 18:16 |
Hieronymous | lambday: great, seems like a good starting point. Thanks. | 18:16 |
@lambday | Hieronymous: not at all :) | 18:17 |
@sukey | New Commit "Fixed the two-sample test notebook with the latest API" to shogun-toolbox/shogun by lambday: https://github.com/shogun-toolbox/shogun/commit/6b09feb2f12c358bbb0ed1d471a07b4501b31139 | 18:52 |
Hieronymous | lambday: OXPHOS is working on the cholseky decomposition. | 19:06 |
Hieronymous | If I had to add something like QR or SVD, any hints on how I should proceed ? | 19:07 |
-!- abhinav [~quassel@103.211.53.167] has quit [Ping timeout: 250 seconds] | 19:07 | |
Hieronymous | lambday: I actually checked OXPHOS's implementation of Cholesky. All of the operations are done using Eigen. | 19:07 |
Hieronymous | Any ideas on how I could begin ? Just general outline any resource or implementation I should model it on. | 19:08 |
Hieronymous | lambday: Like I said, just hints/guidelines :) Thanks. | 19:09 |
@lambday | Hieronymous: you can add other factorization methods :) | 19:09 |
@lambday | wait let me find you one issue on Github | 19:10 |
Hieronymous | Yeah QR/SVD. Actually there is - 2526 | 19:10 |
Hieronymous | #2526 | 19:10 |
@lambday | https://github.com/shogun-toolbox/shogun/issues/2526 | 19:10 |
@lambday | exactly!~ | 19:10 |
Hieronymous | that is. I looked at those, QR or SVD hasn't been attempted by anyone. I was looking at OXPHOS's implementation | 19:10 |
Hieronymous | of Cholesky in linalg. | 19:11 |
@lambday | Hieronymous: yeah so you can add those two | 19:11 |
Hieronymous | I know all operations are done by Eigen anyway. | 19:11 |
Hieronymous | Just that I don't understand how I should actually begin ? I am new. | 19:11 |
@lambday | Hieronymous: that's what we use for our CPU backend | 19:11 |
@lambday | Hieronymous: no worries.. | 19:11 |
@lambday | maybe you can spend some time learning things about Eigen ? | 19:12 |
@lambday | how to do these things properly | 19:12 |
Hieronymous | Actually I did. I installed Eigen and am playing around with the QR module. | 19:12 |
Hieronymous | I will do that then. | 19:12 |
@lambday | and to know what each operation costs.. really important if you ask me | 19:12 |
@lambday | QR can just be another method parallel to cholesky | 19:13 |
@lambday | so you can add the new method by looking and following the same format | 19:13 |
@lambday | should get you started smoothly :) | 19:13 |
@lambday | Hieronymous: it's really important to have an idea about what exactly each operation does - how much additional memory it allocates etc etc.. whenever in doubt, write a small program like this shitty one - https://gist.github.com/lambday/2c67201d6e75c2aa8e10e5eb344df2b4 | 19:16 |
Hieronymous | lambday: Thanks a ton! I will fiddle around with this. | 19:17 |
@lambday | having an idea would help us prevent writing codes that are easy to write but shit when it comes to performance | 19:17 |
Hieronymous | lambday: So all the decomposition methods go into the backend specifically - >LinalgBackendEigen.h ? | 19:18 |
@lambday | Hieronymous: for now, yes | 19:18 |
@lambday | we will eventually add the GPU support | 19:18 |
@lambday | but one thing at a time | 19:18 |
@lambday | :) | 19:19 |
Hieronymous | lambday: I don't think I will be writing anything by hand. Eigen provides QR decomposition via a bunch of different methods. | 19:19 |
Hieronymous | I will have to research on them then. | 19:19 |
Hieronymous | I have gotten to know from the chat logs etc that you're the go to guy for linalg. | 19:19 |
Hieronymous | Hence these many questions. | 19:19 |
@lambday | Hieronymous: yeah.. the suggestion was not only about QR in particular.. but about using any library methods | 19:20 |
@lambday | Hieronymous: well, I am not the guy anymore :) but I'll try to help as much as I can | 19:20 |
Hieronymous | I will definitely research a bit. Although I would leave it in good faith that the guys who built Eigen, did their thorough research | 19:20 |
Hieronymous | And cut out some work for noobs like me. | 19:20 |
@lambday | Hieronymous: if you really want to work on linalg, you should definitely have a chat with Pan (OXPHOS) | 19:21 |
@lambday | collaborate | 19:21 |
Hieronymous | Yeah, Can't seem to catch her. I am more or less here the entire day. Except when I'm sleeping, of course. | 19:21 |
@lambday | it was her project this summer.. she knows it better than I do :) | 19:21 |
Hieronymous | lambday: Thanks! Will do. | 19:22 |
@lambday | Hieronymous: you can tag @OXPHOS on github if you have questions on linalg | 19:22 |
@lambday | since GSoC is over, people (including me) are not online that often :( | 19:23 |
Hieronymous | I will do that on #2526 then. :) | 19:23 |
@lambday | Hieronymous: cool! sounds like a plan! | 19:23 |
@wiking | oh mqn | 19:23 |
Hieronymous | I understand work can eat a lot of time. I actually did my GSoC while I was interning at a diff place | 19:23 |
@lambday | Hieronymous: and thanks for your interest on this.. linalg is an exciting endeavour | 19:23 |
@wiking | we need to fix our 3.3.0 eigen integration Eigen 3.3 released! | 19:23 |
@lambday | Hieronymous: yeha I know | 19:24 |
@lambday | wiking: yo! how you doin' ? :D | 19:24 |
@wiking | jetlag | 19:24 |
@lambday | why I am not surprised | 19:24 |
lisitsyn | HO | 19:24 |
@lambday | wiking: what's the fix required for that? | 19:24 |
@wiking | mmm | 19:24 |
@lambday | lisitsyn: yo!!! | 19:24 |
@wiking | locode hax | 19:24 |
lisitsyn | hea | 19:24 |
@wiking | *code hacks | 19:25 |
@wiking | but yeah i'll have to see now | 19:25 |
@wiking | as it has been released not long ago | 19:25 |
@lambday | wiking: let me know if I can help with that | 19:25 |
@wiking | and wondering which are the fixes they managed to get in | 19:25 |
@lambday | wiking: BTW what happened with our Windows build? | 19:25 |
@wiking | as there were some crazy bugs in 3.3 beta releases with iterators | 19:25 |
@wiking | lambday, need to fix crossval | 19:25 |
@lambday | lisitsyn: how have you been? I've been absent for a long time | 19:26 |
@wiking | that's the major failing test that is failing | 19:26 |
lisitsyn | lambday: me too | 19:26 |
lisitsyn | :D | 19:26 |
@lambday | wiking: that uses Eige! | 19:26 |
@wiking | yeah it's not about 3.3 | 19:26 |
@wiking | as we bundle there the library | 19:26 |
@wiking | i mean eigen | 19:26 |
@lambday | so it always downloads the latest version? | 19:26 |
@lambday | and it breaks the shit out | 19:27 |
@lambday | :/ | 19:27 |
@wiking | noup | 19:27 |
@wiking | when we bundle | 19:27 |
@wiking | it's totally fixed | 19:27 |
@wiking | which is the version being bundled | 19:27 |
@wiking | 3.2.whatever | 19:27 |
@wiking | crossval is broken | 19:27 |
@lambday | lisitsyn: :D I am back to the same fucking job :( | 19:27 |
@lambday | wiking: oh! | 19:27 |
lisitsyn | lambday: how come? | 19:28 |
@lambday | lisitsyn: well, for now.. deciding about applying for PhD | 19:28 |
@lambday | since this job gives me enough personal time, so, good for now | 19:28 |
@lambday | :D | 19:28 |
@lambday | wiking: is there any unit test that I can try to reproduce this error? | 19:29 |
@lambday | anything you noted? | 19:29 |
@wiking | yeah | 19:29 |
@wiking | the crossval unit test | 19:29 |
@lambday | okay | 19:29 |
@wiking | but the error only comes out on windo | 19:29 |
@lambday | I'll have a look then | 19:29 |
@wiking | ws | 19:29 |
@lambday | argh! | 19:30 |
@wiking | an as well | 19:31 |
@wiking | ardkernel is broken on osx | 19:31 |
@wiking | and you did some changes | 19:31 |
@wiking | in gaussian kernels | 19:31 |
@wiking | that totally fucked up the kernel normalizers | 19:32 |
@wiking | :) | 19:32 |
@wiking | so rather fix that part | 19:32 |
@wiking | i think Heiko has opened an issue for that bug | 19:33 |
@lambday | wiking: I see.. yeah I made the distance precomputable | 19:34 |
@lambday | couldn't check on OSX | 19:34 |
@wiking | yep and it does not comply with | 19:34 |
@wiking | normalizers | 19:34 |
@wiking | the normallizer thing | 19:35 |
@wiking | is reproducable anywhere | 19:35 |
@lambday | wiking: shall I look at our buildbots then? | 19:35 |
@lambday | okay | 19:35 |
@wiking | just check for the issue | 19:35 |
@wiking | it'll explain what's the problem | 19:35 |
@lambday | wiking: cool.. let me check | 19:35 |
@wiking | https://github.com/shogun-toolbox/shogun/issues/3504 | 19:36 |
@lambday | https://github.com/shogun-toolbox/shogun/issues/3504 | 19:36 |
@lambday | haha | 19:36 |
@lambday | wait let me read | 19:36 |
@lambday | I am assigning it to myself | 19:36 |
@lambday | will send a fix | 19:36 |
@wiking | k | 19:36 |
@wiking | thnx | 19:36 |
@sukey | Issue #3504 "fix kernel normalisers for GaussianKernel" assigned to: lambday by lambday - https://github.com/shogun-toolbox/shogun/issues/3504 | 19:37 |
@lambday | wiking: when are we planning for 5.1? | 19:38 |
@lambday | any dates? | 19:38 |
@wiking | sooner the better | 19:38 |
@wiking | but i think due to some serious bugs | 19:38 |
@wiking | blocking us for osx release | 19:38 |
@wiking | we are gonna have 5.0.1 | 19:38 |
@wiking | as soon as we have those fixes | 19:38 |
@wiking | if everything goes ok that could be already during this weekend | 19:38 |
@lambday | tell me about our versioning convention again.. x.0.1 is a bugfix release? | 19:39 |
@lambday | x.1 is a major release? | 19:39 |
@wiking | it's not our convension | 19:39 |
@wiking | <major>.<minor>.<patch> | 19:39 |
@wiking | http://semver.org/ | 19:39 |
@lambday | so bugfix is just a patch | 19:40 |
@wiking | yep | 19:40 |
@wiking | but this patch | 19:40 |
@wiking | that you need to do | 19:40 |
-!- travis-ci [~travis-ci@ec2-54-162-82-185.compute-1.amazonaws.com] has joined #shogun | 19:40 | |
travis-ci | it's lambday'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/176453657 | 19:40 |
-!- travis-ci [~travis-ci@ec2-54-162-82-185.compute-1.amazonaws.com] has left #shogun [] | 19:40 | |
@wiking | might need to be under major | 19:40 |
@wiking | because the api should change | 19:40 |
@wiking | as you should remove using friend | 19:41 |
@wiking | i mean of course | 19:41 |
@wiking | depends how you are planning to fix this whole story | 19:41 |
@lambday | gotit | 19:41 |
@lambday | wiking: I guess when we merge my bigtest feature branch, it should also go under a major release.. cause the API changed | 19:43 |
@lambday | but we'll discuss that later | 19:43 |
@wiking | yeah i mean around jan | 19:43 |
@wiking | we'll have 6.0 | 19:43 |
@wiking | but first we wanna have a bugfix release | 19:44 |
@wiking | 5.1 | 19:44 |
@wiking | lambday, add to the milestone https://github.com/shogun-toolbox/shogun/milestone/11 | 19:44 |
@lambday | wiking: cool | 19:44 |
@wiking | and here's the long list of bugs needs to be closed for 5.1 | 19:44 |
@wiking | https://github.com/shogun-toolbox/shogun/milestone/10 | 19:44 |
@lambday | crap I really have been off for a long time :( | 19:45 |
@lambday | looks like the issues I was talking about are already added to the milestone | 19:46 |
@lambday | wiking: did we make c++11 mandatory for 5 release already? | 19:47 |
@wiking | noup | 19:48 |
@wiking | that's 6.0 | 19:48 |
@lambday | should go on Jan release then | 19:48 |
@lambday | yeah | 19:48 |
* lambday going off | 20:20 | |
-!- lambday [6a331336@gateway/web/freenode/ip.106.51.19.54] has quit [] | 20:20 | |
-!- Hieronymous [~sudarshan@122.169.172.64] has quit [Quit: Leaving] | 20:23 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 21:47 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 21:51 | |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has joined #shogun | 22:53 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 22:53 | |
@wiking | HeikoS, new waldeck is good | 23:20 |
@HeikoS | wiking: cool will check | 23:20 |
@wiking | https://play.spotify.com/album/0YxXmh88Sl26u5plhL2cxC?play=true&utm_source=open.spotify.com&utm_medium=open | 23:21 |
@HeikoS | wiking: hahaha :D | 23:30 |
@wiking | HeikoS, here? | 23:48 |
@HeikoS | yeah | 23:48 |
@wiking | HeikoS, so they just relesed | 23:48 |
@wiking | couple of days ago 3.3.0 | 23:48 |
@wiking | eigne | 23:48 |
@wiking | eigen | 23:49 |
@wiking | :))) | 23:49 |
@HeikoS | ay | 23:49 |
@wiking | so yeah | 23:49 |
@wiking | now fixing all that shit | 23:49 |
@HeikoS | with that thing in the changelog? | 23:49 |
@wiking | :D | 23:49 |
@wiking | but the iterator thing | 23:49 |
@wiking | is fixed | 23:49 |
@wiking | thnx god | 23:49 |
@wiking | :) | 23:49 |
@HeikoS | cool | 23:49 |
@HeikoS | and the operator mess? | 23:49 |
@wiking | i think is well | 23:49 |
@HeikoS | cool | 23:49 |
@wiking | but now i have to fix osx | 23:49 |
@HeikoS | great | 23:49 |
@wiking | so i will tell you once that's done | 23:49 |
@HeikoS | I want to write a bit of dev docs | 23:49 |
@wiking | because basically i have the eigen 3.3.0 locally | 23:49 |
@HeikoS | I spend too much time epxlaining to the kids | 23:49 |
@wiking | so i'm doing both at the same time | 23:49 |
@wiking | yeah i see that | 23:50 |
@wiking | :D | 23:50 |
@wiking | let them write it | 23:50 |
@wiking | FAQ | 23:50 |
@HeikoS | no | 23:50 |
@wiking | :) | 23:50 |
@HeikoS | takes too much time | 23:50 |
@HeikoS | proof reading and feedbacking it | 23:50 |
@wiking | kk | 23:50 |
@HeikoS | I will do very minimal | 23:50 |
@HeikoS | since I agree with lisitsyn that docs are controversial as they outdate fast | 23:50 |
@HeikoS | and wrong docs is even worse than no docs ;) | 23:50 |
@HeikoS | but yeah will put some time into that | 23:50 |
lisitsyn | TRUE THAT | 23:50 |
@wiking | :> | 23:50 |
lisitsyn | wiking: you know what? | 23:51 |
lisitsyn | I know you're python fan | 23:51 |
@wiking | :>>>>>>>>>>>>>>>>>>>>>>>> | 23:51 |
lisitsyn | so well | 23:51 |
@wiking | fuck YOU | 23:51 |
@wiking | :D | 23:51 |
@wiking | and fuck alllllllllllllll this z generation | 23:51 |
@wiking | where this is the norm | 23:51 |
@wiking | https://lh3.googleusercontent.com/-7UdlpIo6Qy8/WBrIGUShGcI/AAAAAAAB4iA/4FBjKS-UV9A0O3YuWBqYktnB4TBUCErnwCJoC/w639-h471/2016-11-02.gif | 23:52 |
@wiking | lisitsyn, ^ | 23:52 |
@HeikoS | haha | 23:52 |
@HeikoS | epic | 23:52 |
@HeikoS | wiking: btw today i played a bit with tensorflow | 23:53 |
@wiking | yeah this is like the perfect picture of the state of the art | 23:53 |
@HeikoS | it is so much nicer than theano | 23:53 |
@HeikoS | yet very very slow if you just write down what you think it should be | 23:53 |
@HeikoS | I might start a little mini lib for phd | 23:53 |
@HeikoS | and then learn more about it that way | 23:53 |
@HeikoS | so we can copy andor use good bits | 23:54 |
lisitsyn | wiking: soo | 23:54 |
lisitsyn | the thing is | 23:54 |
lisitsyn | a few weeks ago I wrote a mail here at my job | 23:54 |
lisitsyn | like lets start writing java | 23:54 |
lisitsyn | :D | 23:54 |
@wiking | :DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD | 23:55 |
lisitsyn | it was for team of ~20 | 23:55 |
lisitsyn | and we actually started | 23:55 |
lisitsyn | and now I am constantly having fun | 23:55 |
lisitsyn | wiking: for example a few guys were making a mock | 23:56 |
lisitsyn | because they feel anxious about their code | 23:56 |
lisitsyn | failing in the night | 23:56 |
lisitsyn | because of typo | 23:56 |
lisitsyn | how fucking come we get there? | 23:56 |
lisitsyn | is it any sane :D | 23:56 |
@wiking | yeah man | 23:57 |
@wiking | that is fucking insane | 23:57 |
@wiking | spelling error | 23:57 |
@wiking | and maybe it's just in the catch part of the try block | 23:57 |
@wiking | so it only gets there when there's the exception | 23:57 |
lisitsyn | yeah | 23:57 |
lisitsyn | except NameError | 23:57 |
@wiking | and then suddenly like booom | 23:57 |
lisitsyn | ok ok it wokrs | 23:57 |
lisitsyn | :D | 23:57 |
@wiking | no such variable | 23:57 |
@wiking | :D | 23:57 |
@wiking | because you actually did a spelling mistake or refactor mistake | 23:57 |
@wiking | and did not change the variable part | 23:58 |
@wiking | in the except part of the thing | 23:58 |
@wiking | and then you are like | 23:58 |
@wiking | WTF is this place? | 23:58 |
@wiking | :D | 23:58 |
@wiking | as a friend told me one day | 23:58 |
@wiking | "i'm too old to learn dynamic languages" | 23:58 |
@wiking | :) | 23:58 |
lisitsyn | so I am actually moving from | 23:59 |
lisitsyn | like getting used to python | 23:59 |
@wiking | yeh me too | 23:59 |
@wiking | i loved it | 23:59 |
@wiking | like 3 years ago | 23:59 |
lisitsyn | to having fun of python people | 23:59 |
@wiking | i was like this is great | 23:59 |
@wiking | :) | 23:59 |
@wiking | but now i'm like | 23:59 |
@wiking | are you fucking serious:) | 23:59 |
--- Log closed Thu Nov 17 00:00:01 2016 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!