--- Log opened Wed Feb 22 00:00:35 2017 | ||
-!- travis-ci [~travis-ci@ec2-54-204-179-254.compute-1.amazonaws.com] has joined #shogun | 00:26 | |
travis-ci | it'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/203989223 | 00:26 |
---|---|---|
-!- travis-ci [~travis-ci@ec2-54-204-179-254.compute-1.amazonaws.com] has left #shogun [] | 00:26 | |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has joined #shogun | 01:09 | |
travis-ci | it'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/203993841 | 01:09 |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has left #shogun [] | 01:09 | |
@sukey | Pull Request #3620 "Add time comparison for KNN_KDTREE" synchronized by MikeLing - https://github.com/shogun-toolbox/shogun/pull/3620 | 02:24 |
-!- mikeling [uid89706@gateway/web/irccloud.com/x-cdfbgmlceusqskbo] has joined #shogun | 03:27 | |
-!- lambday [31cf349d@gateway/web/freenode/ip.49.207.52.157] has joined #shogun | 03:37 | |
-!- mode/#shogun [+o lambday] by ChanServ | 03:37 | |
-!- goksinen_ [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Remote host closed the connection] | 03:49 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 04:02 | |
@lambday | wiking: there? | 04:07 |
@wiking | yes | 04:07 |
@wiking | gonna be away | 04:08 |
@wiking | for 10 muns | 04:08 |
@wiking | *mins | 04:08 |
@lambday | wiking: okay.. on travis the build seems inconsistent | 04:08 |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Remote host closed the connection] | 04:09 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 04:09 | |
@sukey | Pull Request #3534 "LinalgRefactor - Cholesky - CPU only" synchronized by OXPHOS - https://github.com/shogun-toolbox/shogun/pull/3534 | 04:32 |
@wiking | lambday, back | 05:26 |
@wiking | lambday, yeah it is | 05:26 |
@wiking | but it is because it's multiprocessing | 05:26 |
@wiking | i mean multithreading | 05:26 |
@wiking | i.e. | 05:26 |
@wiking | so it's not travis' fault | 05:26 |
@wiking | it is actually a bug | 05:26 |
@lambday | wiking: yeah.. but after the fix, I am not being able to reproduce the error in my local.. it works fine.. | 05:41 |
@lambday | wiking: any ideas on how I can reproduce this? | 05:41 |
@wiking | mmm | 05:43 |
@sukey | Pull Request #3635 "LinalgRefactor - Memory Transfer Mutex" synchronized by OXPHOS - https://github.com/shogun-toolbox/shogun/pull/3635 | 05:45 |
@wiking | lambday, managed to reproduce it... but only outside of gdb :) | 05:58 |
@sukey | New Commit "fix a memory leak in multikernel mmd" to shogun-toolbox/shogun by lambday: https://github.com/shogun-toolbox/shogun/commit/26098beb0b05c148f18c13fb920a345f3dbdf3fa | 06:13 |
@lambday | wiking: just ran it once again.. same thing here :( if you can let me know some more details then maybe I can try to fix it | 06:16 |
@lambday | wiking: going to work now... will be back 8-ish hrs later but I'll be online from office to keep track | 06:17 |
@wiking | lambday, ok... but i'm just rewriting a bit that input parser | 06:25 |
@wiking | and parse buffer | 06:25 |
@wiking | as it's really not necessary to use mutexes there | 06:25 |
@wiking | :) | 06:25 |
-!- lambday_ [c40f1708@gateway/web/freenode/ip.196.15.23.8] has joined #shogun | 06:53 | |
-!- mode/#shogun [+o lambday_] by ChanServ | 06:53 | |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has joined #shogun | 07:22 | |
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/204074298 | 07:22 |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has left #shogun [] | 07:22 | |
@lambday_ | hmm looks like I broke a couple of other tests... | 07:40 |
@lambday_ | but the streaming test passed | 07:41 |
@sukey | New Commit "Create TODO.md" to shogun-toolbox/shogun by lambday: https://github.com/shogun-toolbox/shogun/commit/dd3949ab5fd41bad323dce14fc9eb559275a419c | 08:00 |
@wiking | lambday for the future | 08:10 |
@wiking | could you use tickets or project (https://github.com/shogun-toolbox/shogun/projects | 08:10 |
@wiking | for this todo? | 08:10 |
@wiking | plz | 08:10 |
@lambday_ | wiking: sure I'll do that.. | 08:37 |
@wiking | because now this would be shipped with any distribution | 08:40 |
@wiking | and i dont see really the point of that | 08:40 |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has joined #shogun | 08:53 | |
travis-ci | it's Soumyajit De'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/204090022 | 08:53 |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has left #shogun [] | 08:53 | |
@wiking | lambday here? | 08:58 |
@wiking | or lambday_ here? | 08:58 |
@lambday_ | wiking: hey | 08:59 |
@wiking | ok so i was a bit working on some shit with these streaming features | 08:59 |
@wiking | and i just realised something | 08:59 |
@wiking | what the fuck does | 08:59 |
@wiking | streamingFeature.duplicate() | 08:59 |
@wiking | mean | 08:59 |
@wiking | not syntactically | 08:59 |
@wiking | but semantically | 08:59 |
@wiking | what would you expect there? | 09:00 |
@wiking | you have 2 Streaming objects | 09:00 |
@wiking | and then? | 09:00 |
@lambday_ | wiking: another copy of the streaming feature object | 09:00 |
@wiking | yeah but ok | 09:00 |
@wiking | that i understand | 09:00 |
@wiking | but functionally | 09:00 |
@wiking | then what? | 09:00 |
@lambday_ | dunno... they should be able to stream independent of each other | 09:00 |
@wiking | so you have the copy of the thing | 09:00 |
@wiking | and totally independent of each other? | 09:01 |
@wiking | or they are viewing the same stream/ | 09:01 |
@wiking | ? | 09:01 |
@lambday_ | wiking: talking just from semantics, they should be able to share the same matrix (assuming that it's a streaming dense feature object).. | 09:01 |
@lambday_ | just a sec.. brb | 09:02 |
@wiking | yeah the sharing of the same matrix i get | 09:02 |
@wiking | i mean the input they are streaming should be the same | 09:02 |
@wiking | but | 09:02 |
@wiking | do you expect it to just to be a 'shared' view of the stream | 09:02 |
@wiking | or that they are reading the stream totally independent way? | 09:02 |
@wiking | i.e. if A does some read() | 09:03 |
@wiking | B's read() wouldnt move | 09:03 |
@wiking | ? | 09:03 |
@wiking | so they are just different objects over the same stream | 09:03 |
@wiking | or actually | 09:03 |
@wiking | they are reading the streams in a different manner | 09:03 |
@wiking | so basically | 09:04 |
@wiking | x = a.get_streamed_features(1); | 09:04 |
@wiking | y = b.get_streamed_features(1); | 09:04 |
@wiking | x == y? | 09:05 |
@wiking | or y is the element after x | 09:05 |
@lambday_ | back | 09:06 |
@lambday_ | wiking: well, I'd expect x.equals(y), even though x != y... because it creates temporary small matrices.. based on the streaming size | 09:07 |
@wiking | mmm | 09:07 |
@wiking | what do you mean 'temporary small matrices'? | 09:07 |
@wiking | this is a very simple question | 09:07 |
@wiking | is x == y ? | 09:08 |
@wiking | if you b.get_streamed_features(1) | 09:08 |
@wiking | on both objects | 09:08 |
@wiking | where b = a.duplicate() | 09:08 |
@lambday_ | say, we have 10000 features... streaming size is 10.. then it returns a new matrix, copying that 10 | 09:08 |
@lambday_ | wiking: yeah... they are same, value-wise.. | 09:09 |
@wiking | get_streamed_features( num_elements) | 09:09 |
@wiking | so you request the num elements there | 09:09 |
@wiking | so you get N values back | 09:09 |
@wiking | in a CFeatures | 09:09 |
@wiking | ok so they are totally independent streams | 09:09 |
@wiking | ? | 09:09 |
@wiking | over the same input | 09:09 |
@lambday_ | wiking: I am just guessing.. semantically they should be independent.. | 09:10 |
@wiking | i mean | 09:10 |
@wiking | we should decide | 09:10 |
@wiking | because currently | 09:10 |
@wiking | it's just another view | 09:10 |
@lambday_ | on a, I should be able to do a.get_streamed_features(N) and on b, b.get_streamed_features(M) | 09:10 |
@wiking | of the input | 09:10 |
@lambday_ | hmm | 09:11 |
@wiking | ok this needs fixing then | 09:11 |
@wiking | because currently | 09:11 |
@wiking | the mutex-es | 09:11 |
@wiking | are | 09:11 |
@wiking | shared_ptr | 09:11 |
@wiking | namely they are operating over the same fucking | 09:11 |
@wiking | mutex | 09:11 |
@wiking | so there should be an explicit copy of the mutex | 09:11 |
@lambday_ | yeah.. | 09:11 |
@lambday_ | is this duplicate method used anywhere in Shogun? | 09:12 |
@lambday_ | why not call it clone then :/ | 09:12 |
@wiking | this is virtual of CFeatures | 09:12 |
@lambday_ | ah | 09:12 |
@wiking | and btw its not being used anywhere | 09:13 |
@wiking | so | 09:13 |
@wiking | i woudl say | 09:13 |
@wiking | i would not implement this for streaming features? | 09:13 |
@wiking | and even delete the copy ctor of it? | 09:13 |
@wiking | hmm? | 09:13 |
@wiking | just to avoid confusion | 09:13 |
@lambday_ | wiking: maybe that's fine... SG_NOTIMPLEMENTED | 09:13 |
@wiking | yeah just want to explicitly disable copy ctor on these streaming stuff | 09:13 |
@wiking | so that nobody tries to copy it | 09:13 |
@wiking | or if they want | 09:14 |
@wiking | they can | 09:14 |
@wiking | just implement it correctly | 09:14 |
@lambday_ | yeah otherwise the shared_ptr with mutex will get screwed up | 09:14 |
@wiking | yes | 09:14 |
@wiking | although now i'm trying to kill all the mutexes | 09:14 |
@wiking | with atomic | 09:14 |
@wiking | but the story there the same | 09:14 |
@wiking | atomic's copy ctor is = delete | 09:14 |
@wiking | so there you need explicit copying as well | 09:14 |
@lambday_ | so you **have** to delete your copy ctor | 09:14 |
@wiking | yep | 09:15 |
@lambday_ | wiking: I am going for lunch.. will be back in 30 mins | 09:17 |
@wiking | lambday just one question | 09:17 |
@wiking | am i righ tthat if you delete the parent's copy ctor | 09:17 |
@wiking | the derived classes will implicitly have their copy ctor 'deleted' | 09:17 |
@wiking | yeah i think so :) | 09:17 |
@wiking | ok it is | 09:19 |
@wiking | clang is exactly telling me this | 09:19 |
@wiking | perfect | 09:19 |
@wiking | :) | 09:19 |
-!- rushabh [67334b7e@gateway/web/freenode/ip.103.51.75.126] has joined #shogun | 09:41 | |
@lambday_ | wiking: back.. so it is working now? | 10:01 |
@wiking | mmm no | 10:01 |
@wiking | it still needs more fixing | 10:02 |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has joined #shogun | 10:02 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 10:02 | |
@wiking | HeikoS, y0 | 10:03 |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has quit [Ping timeout: 255 seconds] | 10:07 | |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has joined #shogun | 10:53 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 10:53 | |
@sukey | Issue #3652 "investigate XGBoost" karlnapf added label: "development tasks" - https://github.com/shogun-toolbox/shogun/issues/3652 | 11:18 |
@sukey | Issue #3652 "investigate XGBoost" karlnapf added label: "entrance" - https://github.com/shogun-toolbox/shogun/issues/3652 | 11:18 |
@sukey | Issue #3652 "investigate XGBoost" opened by karlnapf - https://github.com/shogun-toolbox/shogun/issues/3652 | 11:18 |
@HeikoS | lisitsyn: jo | 11:23 |
@HeikoS | you around? | 11:23 |
@wiking | HeikoS, my phone is dying | 11:43 |
@wiking | i'm gonna be back in a sec | 11:43 |
@HeikoS | wiking: kk | 11:43 |
mikeling | HeikoS: hi, I had update the KNN notebook and the gist in https://gist.github.com/MikeLing/c57eef21a17c4371e2dcb3f4ce7505eb | 11:51 |
@HeikoS | wiking: https://github.com/libdynd/libdynd | 11:53 |
@HeikoS | you ever saw this? | 11:53 |
@sukey | Pull Request #3620 "Add time comparison for KNN_KDTREE" synchronized by MikeLing - https://github.com/shogun-toolbox/shogun/pull/3620 | 11:53 |
@HeikoS | ah it is c++14 | 11:54 |
@wiking | :> | 11:54 |
mikeling | and I also create pr to shogun-data in https://github.com/shogun-toolbox/shogun-data/pull/134 ;) | 11:54 |
@HeikoS | mikeling: will check soon | 11:58 |
mikeling | HeikoS: thank you! :) | 11:59 |
@sukey | New Commit "fix a memory leak in multikernel mmd" to shogun-toolbox/shogun by karlnapf: https://github.com/shogun-toolbox/shogun/commit/26098beb0b05c148f18c13fb920a345f3dbdf3fa | 12:13 |
@sukey | New Commit "doc fix for resize_vector and unit test" to shogun-toolbox/shogun by karlnapf: https://github.com/shogun-toolbox/shogun/commit/6f95a077aeab0f4569b580cce533a64c77ff0ea7 | 12:20 |
@sukey | New Commit "change index_t casts of size_t to int64_t where safe" to shogun-toolbox/shogun by karlnapf: https://github.com/shogun-toolbox/shogun/commit/7b2b73504306eda5c72db29243f9f4737d908a17 | 12:28 |
CaBa | heya | 12:43 |
@HeikoS | jojo | 12:44 |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has quit [Ping timeout: 255 seconds] | 12:57 | |
-!- lambday_ [c40f1708@gateway/web/freenode/ip.196.15.23.8] has quit [Quit: Page closed] | 13:05 | |
@sukey | Issue #3651 "LibSVM train() error. (with BinaryLabels)"- https://github.com/shogun-toolbox/shogun/issues/3651 | 13:27 |
-!- travis-ci [~travis-ci@ec2-54-224-148-1.compute-1.amazonaws.com] has joined #shogun | 13:29 | |
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/204152352 | 13:29 |
-!- travis-ci [~travis-ci@ec2-54-224-148-1.compute-1.amazonaws.com] has left #shogun [] | 13:29 | |
-!- suhas2go [uid201652@gateway/web/irccloud.com/x-pwjlxepjpzgtegbh] has joined #shogun | 14:09 | |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has joined #shogun | 14:10 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 14:10 | |
@HeikoS | lisitsyn: jojo | 14:10 |
@HeikoS | sonney2k: around? | 14:11 |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has joined #shogun | 14:15 | |
travis-ci | it'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/204154375 | 14:15 |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has left #shogun [] | 14:15 | |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has quit [Ping timeout: 268 seconds] | 14:19 | |
@lambday | back | 14:49 |
@sukey | New Commit "fix segfault in unit tests" to shogun-toolbox/shogun by lambday: https://github.com/shogun-toolbox/shogun/commit/5ba962e86340aa6388d8fcc080454b6855a21211 | 16:26 |
-!- kesslerfrost [~textual@2405:204:103:e41c:c5e5:3ec8:4660:43da] has joined #shogun | 16:37 | |
-!- mikeling [uid89706@gateway/web/irccloud.com/x-cdfbgmlceusqskbo] has quit [Quit: Connection closed for inactivity] | 17:09 | |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has joined #shogun | 17:19 | |
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/204233735 | 17:19 |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has left #shogun [] | 17:19 | |
@sukey | New Commit "make silent underflow failure explicit at runtime" to shogun-toolbox/shogun by karlnapf: https://github.com/shogun-toolbox/shogun/commit/ddf78b11c6f66b6bae559eaac27ae73bdcfd2365 | 17:26 |
-!- kesslerfrost [~textual@2405:204:103:e41c:c5e5:3ec8:4660:43da] has quit [Ping timeout: 240 seconds] | 17:30 | |
-!- kesslerfrost [~textual@49.44.51.104] has joined #shogun | 17:57 | |
@lambday | wiking: hi | 18:12 |
-!- kesslerfrost [~textual@49.44.51.104] has quit [Ping timeout: 260 seconds] | 18:14 | |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has joined #shogun | 18:18 | |
travis-ci | it'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/204258060 | 18:18 |
-!- travis-ci [~travis-ci@ec2-54-196-155-41.compute-1.amazonaws.com] has left #shogun [] | 18:18 | |
-!- kesslerfrost [~textual@49.44.51.104] has joined #shogun | 18:20 | |
CaBa | all that subsetting logic... it is nestable, right? if i apply subsets to my features, i can plug them into the usual machinery like x-val and it will be able to subset it further, right? | 19:09 |
-!- rushabh [67334b7e@gateway/web/freenode/ip.103.51.75.126] has quit [Quit: Page closed] | 19:12 | |
-!- kesslerfrost [~textual@49.44.51.104] has quit [Quit: kesslerfrost] | 20:43 | |
@sukey | New Commit "fixed memory leak in statistical hypothesis test unit tests" to shogun-toolbox/shogun by lambday: https://github.com/shogun-toolbox/shogun/commit/49f28650bf583effbb78bb8be96d2dca84c9f589 | 20:55 |
-!- lambday [31cf349d@gateway/web/freenode/ip.49.207.52.157] has quit [] | 20:56 | |
Saurabh7_ | CaBa: yes it should | 20:57 |
-!- travis-ci [~travis-ci@ec2-54-204-179-254.compute-1.amazonaws.com] has joined #shogun | 21:47 | |
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/204334009 | 21:47 |
-!- travis-ci [~travis-ci@ec2-54-204-179-254.compute-1.amazonaws.com] has left #shogun [] | 21:47 | |
--- Log closed Thu Feb 23 00:00:37 2017 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!