--- Log opened Wed Nov 02 00:00:58 2016 | ||
-!- travis-ci [~travis-ci@ec2-54-205-74-4.compute-1.amazonaws.com] has joined #shogun | 00:10 | |
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/172443639 | 00:10 |
---|---|---|
-!- travis-ci [~travis-ci@ec2-54-205-74-4.compute-1.amazonaws.com] has left #shogun [] | 00:10 | |
-!- mizari [~mizari@95-174-213-100.nts.su] has joined #shogun | 04:09 | |
-!- adi__ [31cf3b3b@gateway/web/freenode/ip.49.207.59.59] has joined #shogun | 05:02 | |
-!- adi__ [31cf3b3b@gateway/web/freenode/ip.49.207.59.59] has quit [Ping timeout: 260 seconds] | 05:32 | |
-!- praisethemoon [~praisethe@41.226.248.123] has joined #shogun | 08:37 | |
-!- HeikoS [~heiko@eduroam-int-pat-8-38.ucl.ac.uk] has joined #shogun | 11:57 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 11:57 | |
@HeikoS | wiking: jojo | 12:01 |
@wiking | HeikoS, hi | 13:38 |
@HeikoS | wiking: jo | 14:27 |
@HeikoS | any updates? | 14:28 |
@wiking | hi | 14:28 |
@wiking | on? | 14:28 |
@HeikoS | release | 14:28 |
@wiking | yeah msged u | 14:28 |
@wiking | did some stuff | 14:28 |
@HeikoS | the windows yeah | 14:28 |
@HeikoS | so infinite loop is not coming from openmp | 14:28 |
@HeikoS | baaad ;) | 14:28 |
@HeikoS | I will file another 5.1 issue I guess | 14:28 |
@HeikoS | log is where again? | 14:29 |
@wiking | https://ci.appveyor.com/project/vigsterkr/shogun/history | 14:30 |
@HeikoS | thanks | 14:32 |
@HeikoS | wiking: so anything I can help with the release? | 14:33 |
@wiking | dunno man | 14:33 |
@HeikoS | otherwise I will polish the wiki and website more | 14:33 |
@wiking | how about the | 14:33 |
@wiking | notebooks? | 14:33 |
@HeikoS | ah yeah | 14:33 |
@HeikoS | I have fixed more | 14:33 |
@HeikoS | let me push | 14:33 |
@wiking | ok push and then i re-run the generator | 14:34 |
@HeikoS | wiking: ah I already pushed that yesterday | 14:35 |
@wiking | ok lemme regen | 14:35 |
@wiking | lets seee | 14:35 |
@HeikoS | Intro | 14:35 |
@HeikoS | is fixed | 14:35 |
@HeikoS | the other ones: I dont get error message locally | 14:35 |
@HeikoS | so maybe its worth that I do this thing that the ntoebooks are executed as python code | 14:36 |
@wiking | pip install -U matplotlib | 14:36 |
@wiking | and you will get them | 14:36 |
@wiking | it's all about scatter() | 14:36 |
@HeikoS | mmh | 14:36 |
@HeikoS | weird because I had a scatter problem in the Intro one | 14:36 |
@HeikoS | but the others went fine | 14:37 |
@HeikoS | can i force the nighly default? | 14:37 |
@wiking | matplotlib (1.4.2) | 14:37 |
@wiking | this is on the nightly_default | 14:37 |
@wiking | you can force anything | 14:38 |
@wiking | but i've already forced it | 14:38 |
@HeikoS | ok :) | 14:38 |
@wiking | http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/6 | 14:38 |
@HeikoS | wiking: saw that sanuj got his blog post on the google blog? | 14:41 |
@wiking | yep | 14:42 |
@wiking | but is it already posted | 14:43 |
@wiking | ? | 14:43 |
@wiking | or is it on the works? | 14:43 |
@HeikoS | they posted 3/10 | 14:43 |
@HeikoS | his is 8 or so | 14:43 |
@HeikoS | pretty cool | 14:43 |
@HeikoS | I asked the others to write one as well would be cool | 14:43 |
@wiking | yep saw it | 14:43 |
@wiking | btw shouldn't we write to Gunnar? | 14:43 |
@HeikoS | yes | 14:44 |
@HeikoS | but lets release first | 14:44 |
@HeikoS | all the exposure to the old website hurts me :) | 14:45 |
@HeikoS | I mean in reaching out to people to talk about it | 14:45 |
@HeikoS | and then they come visit this old shiat | 14:45 |
@wiking | :> | 14:45 |
@wiking | lets see | 14:45 |
@HeikoS | I already sent him the notes we made | 14:45 |
@HeikoS | but we shoul dhave a hangout I think | 14:45 |
@HeikoS | shall I ask about that | 14:45 |
@HeikoS | wiking: email sent :) | 14:46 |
@wiking | sure | 14:46 |
@wiking | :+1: | 14:47 |
@wiking | HeikoS, more problems | 14:50 |
@wiking | :))) | 14:50 |
@wiking | http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/6/steps/notebooks/logs/stdio | 14:50 |
@wiking | [NbConvertApp] Converting notebook /home/buildslave/nightly_default/build/doc/ipython-notebooks/multiclass/multiclass_reduction.ipynb to html | 14:50 |
@wiking | [NbConvertApp] Converting notebook /home/buildslave/nightly_default/build/doc/ipython-notebooks/multiclass/KNN.ipynb to html | 14:50 |
@wiking | [NbConvertApp] Converting notebook /home/buildslave/nightly_default/build/doc/ipython-notebooks/multiclass/Tree/DecisionTrees.ipynb to html | 14:50 |
@wiking | [NbConvertApp] Converting notebook /home/buildslave/nightly_default/build/doc/ipython-notebooks/multiclass/Tree/TreeEnsemble.ipynb to html | 14:50 |
@wiking | [NbConvertApp] Converting notebook /home/buildslave/nightly_default/build/doc/ipython-notebooks/multiclass/naive_bayes.ipynb to html | 14:50 |
@HeikoS | ok checking | 14:51 |
@wiking | mmm | 14:51 |
@wiking | wait | 14:51 |
@wiking | there is something very wrong | 14:51 |
@wiking | ... | 14:51 |
@HeikoS | I was searching for "error |" | 14:51 |
@wiking | suddenly all of them are failing | 14:51 |
@wiking | RuntimeError: Kernel died before replying to kernel_info | 14:51 |
@wiking | :( | 14:51 |
@HeikoS | yeah | 14:51 |
@HeikoS | I mean we have to convert them to python and execute that first | 14:51 |
@wiking | i might need to upgrade jupyter | 14:51 |
@wiking | nono | 14:51 |
@wiking | this is something else | 14:52 |
@HeikoS | ok? | 14:52 |
@wiking | lemme look at it | 14:52 |
@wiking | in the meanwhile | 14:52 |
@wiking | you should find out how to convert to pythno | 14:52 |
@HeikoS | ok checking | 14:52 |
@HeikoS | wiking: this guy right? | 14:57 |
@wiking | which? | 14:57 |
@HeikoS | https://github.com/shogun-toolbox/infra/blob/master/buildbot/scripts/generate-notebooks.sh | 14:57 |
@wiking | what's with it? | 14:57 |
@HeikoS | I modify that | 14:57 |
@wiking | ah you can yeah | 14:58 |
@HeikoS | to execute python code first | 14:58 |
@HeikoS | or rather have another script | 14:58 |
@HeikoS | might be better | 14:58 |
@HeikoS | one execute_notebooks.sh | 14:58 |
@HeikoS | one generate_notebooks.sh | 14:58 |
@HeikoS | or? | 14:58 |
@HeikoS | all in one? | 14:58 |
@wiking | all in one | 14:58 |
@HeikoS | kk | 14:58 |
@wiking | AttributeError: 'AxesSubplot' object has no attribute 'set_prop_cycle' | 15:02 |
@wiking | (in the regression notebook) | 15:02 |
@wiking | should i just fucking downgrade | 15:02 |
@wiking | the matplotlib? | 15:02 |
@wiking | HeikoS, i think all the notebooks are failing on this | 15:14 |
@wiking | too old matplotlib commands | 15:14 |
@wiking | and i actually wonder what is your ubuntu version | 15:14 |
@wiking | that you dont get these errors | 15:14 |
@wiking | ... | 15:14 |
@HeikoS | yeah downgrade for now maybe? | 15:14 |
@wiking | as this is running on a default debian jessie | 15:15 |
@wiking | pacakge | 15:15 |
@HeikoS | I have two upgrades | 15:15 |
@HeikoS | so lots of shit on my system | 15:15 |
@HeikoS | twice lts upgrade | 15:15 |
@HeikoS | not clean | 15:15 |
@wiking | yeah but what is your matplotlib version? | 15:15 |
@wiking | yeah but still | 15:15 |
@wiking | what the fuck is your matplotlib version? | 15:15 |
@HeikoS | 1.5.3 | 15:15 |
@wiking | 1.3.1-1ubuntu5 | 15:16 |
@wiking | hahah lol | 15:16 |
@wiking | then maybe i should UPGRADE | 15:16 |
@wiking | not downgrade | 15:16 |
@HeikoS | just upgraded | 15:16 |
@HeikoS | haha | 15:16 |
@wiking | since we have 1.4.2 | 15:16 |
@wiking | ah ok | 15:16 |
@HeikoS | let me run something | 15:16 |
@HeikoS | which noteobok fails | 15:16 |
@wiking | well then try regression | 15:16 |
@HeikoS | since I just changed things | 15:16 |
@HeikoS | k 1sec | 15:16 |
@wiking | but yeah | 15:16 |
@wiking | more than 50% are failing | 15:16 |
@wiking | so you have a 50% chance to open a shitty one anyways | 15:16 |
@HeikoS | worked | 15:18 |
@HeikoS | regression | 15:18 |
@HeikoS | all cells work | 15:18 |
@wiking | ok so should i upgrade then? | 15:19 |
@wiking | to 1.5.3 | 15:19 |
@wiking | ? | 15:19 |
@HeikoS | yeah maybe try | 15:20 |
@HeikoS | is that reversible easily? | 15:20 |
@wiking | well | 15:21 |
@wiking | it's fucking computers | 15:21 |
@wiking | everything is reversible | 15:21 |
@wiking | :) | 15:21 |
@HeikoS | then go for it | 15:21 |
@HeikoS | I mean it is better to have this updated anyways | 15:21 |
@wiking | ok done | 15:22 |
@wiking | do you have the python command/ | 15:22 |
@wiking | i guess --to python | 15:22 |
@HeikoS | yeah | 15:25 |
@HeikoS | but its tricly | 15:25 |
@HeikoS | just figuring that out | 15:25 |
@HeikoS | because we use %inline stuff | 15:25 |
@wiking | those inlines | 15:25 |
@wiking | are discouraged to use no? | 15:26 |
@HeikoS | nono | 15:26 |
@HeikoS | it is just executing the python file needs to be done with ipython as well | 15:27 |
@HeikoS | so what works is | 15:27 |
@HeikoS | find $1 -type f -name '*.ipynb' | xargs -P $2 -I{} jupyter nbconvert --to script '{}' | 15:27 |
@wiking | ok rerunning | 15:27 |
@wiking | shogun-buildbot_: force build --branch=develop 'nightly_default' | 15:27 |
shogun-buildbot_ | build #7 forced | 15:27 |
shogun-buildbot_ | I'll give a shout when the build finishes | 15:27 |
@HeikoS | wiking: and then need to do something about the graphics, so they are not displayed/blocking | 15:28 |
@HeikoS | wiking: https://pypi.python.org/pypi/runipy | 15:29 |
@wiking | mmm | 15:29 |
@wiking | but yeah | 15:30 |
@wiking | this is not good | 15:30 |
@wiking | coz we have jupyer | 15:30 |
@HeikoS | yeah | 15:30 |
@HeikoS | though does what we want | 15:30 |
@HeikoS | i.e. runs the cells nicely, showing error output | 15:31 |
@HeikoS | wiking: it is maintained | 15:32 |
@wiking | shocking | 15:32 |
@wiking | :) | 15:32 |
@HeikoS | hahahha | 15:32 |
@HeikoS | ah man | 15:32 |
@HeikoS | https://github.com/paulgb/runipy/issues/122 | 15:32 |
@HeikoS | read this | 15:33 |
@HeikoS | lets not use it | 15:33 |
-!- CaBa [~Diu7saig@unaffiliated/caba] has quit [Remote host closed the connection] | 15:34 | |
@wiking | HeikoS, try gmm locally | 15:36 |
@wiking | as it seems to be failing | 15:36 |
@HeikoS | yep | 15:37 |
@HeikoS | Ill fix | 15:37 |
@HeikoS | and push | 15:37 |
@HeikoS | wiking: pushed | 15:38 |
@wiking | k | 15:38 |
@wiking | lets see what's with the others | 15:38 |
@wiking | i wonder why no notification now about pushing to repo | 15:39 |
@wiking | ... | 15:39 |
-!- mizari [~mizari@95-174-213-100.nts.su] has quit [Ping timeout: 260 seconds] | 15:41 | |
@HeikoS | wiking: sooo | 15:47 |
@HeikoS | we can add this guy | 15:47 |
@HeikoS | --allow-errors | 15:47 |
@HeikoS | which is good | 15:47 |
@wiking | what is that | 15:47 |
@wiking | ? | 15:47 |
@HeikoS | allow cells to fail | 15:47 |
@HeikoS | but actually | 15:48 |
@HeikoS | not good | 15:48 |
@HeikoS | since we wanna know | 15:48 |
@wiking | :) | 15:48 |
@HeikoS | ah man | 15:48 |
@HeikoS | these tools | 15:48 |
@HeikoS | why is there no option to print the failing cell output | 15:48 |
@wiking | dunno | 15:48 |
@wiking | you know | 15:49 |
@wiking | opensource | 15:49 |
@wiking | if you dont like it send a patch | 15:49 |
@wiking | :) | 15:49 |
@HeikoS | haha | 15:49 |
@HeikoS | we could actually do this: | 15:49 |
@HeikoS | we put the flag | 15:49 |
@HeikoS | and then upload to some place | 15:49 |
@HeikoS | and then read error there | 15:49 |
@HeikoS | haha | 15:49 |
@wiking | yeah no thanku | 15:50 |
@wiking | [10142929.139176] python[25657]: segfault at 0 ip 00007f962d7e0844 sp 00007fff189936d0 error 4 in libshogun.so.17.2[7f962d1bc000+1988000] | 15:50 |
@wiking | theser are the beautiful ones | 15:50 |
@wiking | :) | 15:50 |
@wiking | i have a feeling that KNN is segfaulting | 15:51 |
@wiking | can u actually test that locally? | 15:51 |
@wiking | and pca_notebook.ipynb | 15:51 |
@wiking | is definitely failing | 15:51 |
@HeikoS | trying both | 15:52 |
@wiking | fgm just failed as well | 15:54 |
@wiking | hahahaha | 15:54 |
@wiking | and tapkee now | 15:54 |
@wiking | which worked before | 15:54 |
lisitsyn | tap kee | 15:54 |
@wiking | prior the matplotlib update | 15:54 |
@HeikoS | pca fixed | 15:54 |
lisitsyn | did I fucked up something? | 15:55 |
@wiking | lisitsyn, as alwasy :D | 15:55 |
@wiking | no | 15:55 |
@wiking | things are messy | 15:55 |
@HeikoS | knn segfaults | 15:55 |
@wiking | HeikoS, yep i thought so | 15:55 |
@HeikoS | uh | 15:55 |
@HeikoS | wiking: ill have a look | 15:57 |
@wiking | k | 15:57 |
@HeikoS | but first want to get error message | 15:57 |
@wiking | check tapkee | 15:57 |
@wiking | that should be an easy fix | 15:57 |
@wiking | since it's just some matplotlib shit | 15:57 |
@wiking | as it worked before the upgrade | 15:57 |
@HeikoS | wiking: works for me | 15:57 |
lisitsyn | how come tapkee + matplotlib? | 15:57 |
@wiking | HeikoS, fuck i dont understand | 15:57 |
@wiking | lisitsyn, old matplotlib | 15:58 |
@wiking | :) | 15:58 |
@HeikoS | wiking: so, it seems that nbconvert should print the error | 15:59 |
@HeikoS | https://github.com/jupyter/nbconvert/pull/120 | 15:59 |
@HeikoS | I mean if a cell fails | 15:59 |
@HeikoS | the way you call it | 15:59 |
@wiking | which version? | 15:59 |
@HeikoS | thjs PR added that | 15:59 |
@HeikoS | sept 2015 | 15:59 |
@HeikoS | checking | 15:59 |
@HeikoS | 4.1 milestone | 16:00 |
@wiking | jupyter (1.0.0) | 16:00 |
@wiking | jupyter-client (4.2.2) | 16:00 |
@wiking | jupyter-console (4.1.1) | 16:00 |
@wiking | jupyter-core (4.1.0) | 16:00 |
@HeikoS | wiking: so if I put an errorneous python statement in notebooks and run it using the jupyter-nbconvert command, I locally get a print in stderr or the thing that failed | 16:01 |
@HeikoS | CellExecutionError: An error occurred while executing the following cell: | 16:01 |
@HeikoS | ------------------ | 16:01 |
@HeikoS | assert False | 16:01 |
@HeikoS | ------------------ | 16:01 |
@HeikoS | AssertionError: | 16:01 |
@wiking | ok tell me the numbers | 16:01 |
@wiking | (version numbers) you have locally | 16:02 |
@wiking | i can pip upgrade then | 16:02 |
@HeikoS | 4.2 | 16:02 |
@wiking | fuck this fucking shi | 16:02 |
@wiking | but which | 16:02 |
@HeikoS | jupyter-nbconvert | 16:02 |
@wiking | of what | 16:02 |
@wiking | fuck | 16:02 |
@wiking | i mean | 16:02 |
@wiking | pip list|grep jupyter | 16:02 |
@wiking | plz | 16:02 |
@HeikoS | jupyter (1.0.0) | 16:02 |
@HeikoS | jupyter-client (4.4.0) | 16:02 |
@HeikoS | jupyter-console (5.0.0) | 16:02 |
@HeikoS | jupyter-core (4.2.0) | 16:02 |
@wiking | ok | 16:02 |
@wiking | upgrading | 16:02 |
@HeikoS | let me know when there is a new log | 16:03 |
@HeikoS | from log, I can usually fix things without running locally | 16:03 |
@wiking | tomorrow | 16:03 |
@HeikoS | cant force it to be there sooner? | 16:03 |
shogun-buildbot_ | Hey! build nightly_default #7 is complete: Exception [exception interrupted] | 16:03 |
shogun-buildbot_ | Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/7 | 16:03 |
@wiking | no | 16:04 |
@wiking | :D | 16:04 |
@wiking | shogun-buildbot_: force build --branch=develop 'nightly_default' | 16:04 |
shogun-buildbot_ | The build has been queued, I'll give a shout when it starts | 16:05 |
@wiking | ok this will have the new output | 16:05 |
@wiking | as just upgraded the shit | 16:05 |
@wiking | but will take some time | 16:05 |
@wiking | coz other bots got triggered in the meanwhile | 16:05 |
shogun-buildbot_ | build #8 forced | 16:06 |
shogun-buildbot_ | I'll give a shout when the build finishes | 16:06 |
@wiking | HeikoS, ^ | 16:06 |
-!- HeikoS [~heiko@eduroam-int-pat-8-38.ucl.ac.uk] has quit [Ping timeout: 256 seconds] | 16:08 | |
-!- HeikoS [~heiko@eduroam-int-pat-8-38.ucl.ac.uk] has joined #shogun | 16:15 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 16:15 | |
@wiking | HeikoS, http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/8/steps/notebooks/logs/stdio | 16:18 |
-!- HeikoS [~heiko@eduroam-int-pat-8-38.ucl.ac.uk] has quit [Ping timeout: 250 seconds] | 16:31 | |
-!- HeikoS [~heiko@eduroam-int-pat-8-38.ucl.ac.uk] has joined #shogun | 16:37 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 16:37 | |
@HeikoS | cool! | 16:37 |
@HeikoS | wiking: so | 16:37 |
@wiking | so there's the error | 16:37 |
@wiking | from fgb | 16:37 |
@wiking | *fgm | 16:37 |
@wiking | but yeah we are almost there | 16:38 |
-!- HeikoS [~heiko@eduroam-int-pat-8-38.ucl.ac.uk] has quit [Ping timeout: 260 seconds] | 16:41 | |
shogun-buildbot_ | Hey! build nightly_default #8 is complete: Failure [failed notebooks] | 16:45 |
shogun-buildbot_ | Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/8 | 16:45 |
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has joined #shogun | 16:55 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 16:55 | |
@wiking | HeikoS, gimme admin rights to the #shogun channel via ChanServ thnx | 16:58 |
@HeikoS | wiking: you dont have? | 16:58 |
@wiking | noup | 16:58 |
@HeikoS | wiking: can you send me the command for that in pm | 16:58 |
@HeikoS | wiking: fixed gp notebook | 16:58 |
@HeikoS | pushing in a sec | 16:59 |
@HeikoS | and fgm is fixed as well | 16:59 |
@HeikoS | opencv I will do in a bit | 16:59 |
-!- mode/#shogun [+o sukey] by ChanServ | 17:00 | |
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has quit [Ping timeout: 250 seconds] | 17:04 | |
-!- shogun-buildbot_ [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 17:19 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 17:19 | |
@wiking | shogun-buildbot_: force build --branch=develop 'nightly_default' | 17:19 |
@wiking | shogun-buildbot: force build --branch=develop 'nightly_default' | 17:19 |
shogun-buildbot | The build has been queued, I'll give a shout when it starts | 17:20 |
-!- HeikoS [~heiko@eduroam-int-pat-8-38.ucl.ac.uk] has joined #shogun | 17:20 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 17:20 | |
@HeikoS | shogun-buildbot: force build --branch=develop 'nightly_default' | 17:21 |
shogun-buildbot | The build has been queued, I'll give a shout when it starts | 17:21 |
@wiking | hahaha | 17:23 |
@wiking | i did that already | 17:23 |
@wiking | we are waiting for other tasks now | 17:23 |
shogun-buildbot | build #9 forced | 17:25 |
shogun-buildbot | I'll give a shout when the build finishes | 17:25 |
@HeikoS | wiking: cancel them all | 17:27 |
@wiking | ? | 17:27 |
@HeikoS | and restart with my latest push | 17:27 |
@HeikoS | had a mistake | 17:27 |
@HeikoS | notebook will fail | 17:27 |
@HeikoS | gp | 17:27 |
@HeikoS | the latest one should only have the opencv failing | 17:28 |
shogun-buildbot | Hey! build nightly_default #9 is complete: Exception [exception interrupted] | 17:28 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/9 | 17:28 |
@HeikoS | thx | 17:28 |
@wiking | shogun-buildbot_: force build --branch=develop 'nightly_default' | 17:28 |
@wiking | shogun-buildbot: force build --branch=develop 'nightly_default' | 17:28 |
@HeikoS | so thats it | 17:28 |
shogun-buildbot | The build has been queued, I'll give a shout when it starts | 17:28 |
@HeikoS | I can look into the opencv but I guess this has low priority? | 17:29 |
@HeikoS | wiking: how is the rest? | 17:29 |
@wiking | nono | 17:30 |
@wiking | you can leave the opencv | 17:30 |
@wiking | knn | 17:30 |
@HeikoS | ah yes | 17:30 |
@wiking | that has a segfault | 17:30 |
@wiking | :) | 17:30 |
@HeikoS | sorry | 17:30 |
@HeikoS | checking | 17:30 |
@HeikoS | haha | 17:30 |
@HeikoS | man so bad :) | 17:30 |
@wiking | :> | 17:30 |
@wiking | it's worth to do it | 17:30 |
@wiking | this way things gets fixed :d | 17:30 |
@HeikoS | yeah I agree, checkign it atm | 17:34 |
@HeikoS | wiking: hows the pipeline coming? | 17:34 |
@wiking | dont get me started | 17:34 |
@HeikoS | man knn code is a f*** | 17:50 |
@HeikoS | Saurabh7: ^ | 17:50 |
@HeikoS | want to clean this up? :) | 17:50 |
-!- praisethemoon [~praisethe@41.226.248.123] has quit [Ping timeout: 268 seconds] | 18:11 | |
-!- HeikoS [~heiko@eduroam-int-pat-8-38.ucl.ac.uk] has quit [Ping timeout: 256 seconds] | 18:14 | |
shogun-buildbot | build #10 forced | 18:15 |
shogun-buildbot | I'll give a shout when the build finishes | 18:15 |
shogun-buildbot | Hey! build nightly_default #10 is complete: Failure [failed notebooks] | 18:25 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/10 | 18:25 |
@wiking | shogun-buildbot: force build --branch=develop 'nightly_default' | 18:27 |
shogun-buildbot | The build has been queued, I'll give a shout when it starts | 18:27 |
shogun-buildbot | build #11 forced | 18:46 |
shogun-buildbot | I'll give a shout when the build finishes | 18:46 |
shogun-buildbot | Hey! build nightly_default #11 is complete: Failure [failed notebooks] | 19:23 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/11 | 19:23 |
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has joined #shogun | 20:56 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 20:56 | |
@HeikoS | wiking: jo! | 21:49 |
@HeikoS | lisitsyn: jo! | 21:49 |
@HeikoS | lisitsyn: I remembered you said covertree KNN was buggy? | 21:49 |
@HeikoS | is that true? | 21:50 |
@HeikoS | wiking: think I fixed the KNN issue | 21:52 |
@HeikoS | I used gdb in fact to debug shogun from ipython notebook :) | 21:53 |
@wiking | \o/ | 22:00 |
@wiking | yeey | 22:00 |
@wiking | pushed? | 22:00 |
@wiking | now it's only | 22:00 |
@wiking | scene classification | 22:01 |
@wiking | and knn is failing | 22:01 |
@wiking | if we could fix knn that'd be great | 22:01 |
@HeikoS | knn is fixed | 22:03 |
@HeikoS | not yet pushed | 22:03 |
@HeikoS | but i have local fix | 22:03 |
@HeikoS | just running unit and integration tests to be sure | 22:03 |
@HeikoS | wiking: so 5 mins | 22:03 |
@wiking | kk | 22:03 |
@wiking | i'm trying to fix that stupid job | 22:03 |
@wiking | that it doesn't stop | 22:03 |
@wiking | if one or more notebooks are failing | 22:04 |
@HeikoS | wiking: it is weird with the notebook | 22:04 |
@HeikoS | that must have been failing for a while | 22:04 |
@wiking | yeah | 22:04 |
@HeikoS | as this definitely was a bug in KNN | 22:04 |
@HeikoS | but not in a recent addition | 22:04 |
@HeikoS | general bug | 22:04 |
@wiking | :) | 22:04 |
@wiking | lol | 22:04 |
@wiking | interesting | 22:04 |
@HeikoS | might have to do with the refactoring of Euclidean distance | 22:04 |
@HeikoS | Actually, I should check that I guess | 22:04 |
@HeikoS | but I know why it breaks | 22:04 |
@HeikoS | and fixed it, so all good | 22:05 |
@HeikoS | man that class | 22:05 |
@HeikoS | Knn.cpp | 22:05 |
@HeikoS | really needs to be sorted out | 22:05 |
@HeikoS | it is SO bad | 22:05 |
@HeikoS | and old | 22:05 |
@wiking | :) | 22:06 |
@HeikoS | wiking: oh and btw | 22:06 |
@HeikoS | this sg_progress is quite nice | 22:06 |
@wiking | what? | 22:06 |
@wiking | ahhahah | 22:06 |
@wiking | but man that's printf | 22:06 |
@HeikoS | as when I press ctrl-c in python | 22:06 |
@HeikoS | signal sorry | 22:07 |
@HeikoS | CSignal | 22:07 |
@HeikoS | it asks whether it should break from the loop | 22:07 |
@HeikoS | that is very nice | 22:07 |
@HeikoS | as program doesnt die | 22:07 |
@HeikoS | just exists the loop | 22:07 |
@wiking | yep | 22:07 |
@wiking | we should generally have that implemented everywhere | 22:07 |
@wiking | i.e. stop convergence | 22:07 |
@HeikoS | yep | 22:07 |
@wiking | but still have the model | 22:07 |
@HeikoS | really good | 22:07 |
@HeikoS | the progress would be cool to be iuntegrated into ipython | 22:08 |
@wiking | ah but | 22:08 |
@wiking | for progress | 22:08 |
@wiking | i would use something like reactivecpp | 22:08 |
@HeikoS | wiking: https://github.com/tqdm/tqdm | 22:08 |
@wiking | https://github.com/Reactive-Extensions/RxCpp | 22:08 |
@HeikoS | connect with this | 22:08 |
@wiking | mmm | 22:08 |
@wiking | lemme see wtf is this | 22:08 |
@HeikoS | as progress bars are mostly useful in interactive sessions | 22:08 |
@wiking | yep | 22:09 |
@wiking | but would be cool | 22:10 |
@wiking | that you can forward this | 22:10 |
@wiking | to anywhere | 22:10 |
@wiking | like not only your STDOUT | 22:10 |
@wiking | but like tcp or whatever | 22:10 |
@wiking | right? | 22:10 |
@wiking | btw with Greg i had the idea | 22:10 |
@wiking | to even be able to publish | 22:10 |
@wiking | the state of the model | 22:10 |
@wiking | i.e. to be able to monitor | 22:10 |
@wiking | w | 22:10 |
@wiking | how does it change iteration by iteration | 22:10 |
@HeikoS | yep | 22:12 |
@HeikoS | great | 22:12 |
@HeikoS | I guess first step is to forward | 22:12 |
@HeikoS | that seems easy | 22:12 |
@HeikoS | maybe comment in the issue I just created :) | 22:12 |
@wiking | HeikoS, http://www.thinkingparallel.com/2007/06/29/breaking-out-of-loops-in-openmp/ | 22:13 |
@HeikoS | wiking: shall we still use openmp? | 22:14 |
@HeikoS | since were moving to c++11 in 6.0? | 22:14 |
@wiking | yes yes | 22:14 |
@wiking | you are not going to implement | 22:14 |
@wiking | a parallel for | 22:14 |
@wiking | with std::thread | 22:14 |
@wiking | :) | 22:14 |
@HeikoS | kk | 22:14 |
@HeikoS | ok ill push now | 22:14 |
@wiking | cool | 22:15 |
@HeikoS | wiking: force da build | 22:15 |
@wiking | just a sec | 22:15 |
@wiking | wanna try to make it sure | 22:15 |
@wiking | somehow | 22:15 |
@wiking | that it still uploads the fucking notebooks | 22:15 |
@wiking | even if there's only 1 fail | 22:15 |
@HeikoS | kk | 22:16 |
@wiking | dont really understand why haltOnFailure=False | 22:16 |
@wiking | doesn't fucking work | 22:16 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 22:28 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 22:29 | |
@wiking | shogun-buildbot: force build --branch=develop 'nightly_default' | 22:30 |
shogun-buildbot | The build has been queued, I'll give a shout when it starts | 22:30 |
@wiking | lets see | 22:30 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Client Quit] | 22:33 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 22:33 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Client Quit] | 22:34 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 22:34 | |
@wiking | lisitsyn HeikoS | 22:53 |
@wiking | so one thing :) | 22:53 |
@wiking | we have this nice | 22:54 |
@wiking | modshogun/doxygen_xml/index.xml | 22:54 |
@wiking | that generates modshogun_doxygen.i | 22:54 |
@wiking | sonney2k_, around? | 23:01 |
@wiking | :) | 23:01 |
@HeikoS | dont know anything about that | 23:01 |
@wiking | hehehe | 23:02 |
@wiking | i wonder why docstrings are generated for everything but ruby | 23:02 |
@wiking | :) | 23:02 |
@HeikoS | ah I see :) | 23:02 |
@wiking | i dont even know that those docstrings could be used from java for example | 23:02 |
@wiking | if not t | 23:02 |
@wiking | then we are generating them | 23:02 |
@wiking | for no use | 23:02 |
@HeikoS | I dont know | 23:03 |
@HeikoS | I guess | 23:04 |
@HeikoS | this is another issue for 5.1 :) | 23:04 |
@HeikoS | Ill create it | 23:04 |
@wiking | HeikoS, btw | 23:04 |
@wiking | what's the naming for python generated tests? | 23:04 |
@wiking | as i believe we have currently a problem | 23:04 |
@wiking | in the modular interface tests | 23:04 |
@wiking | :)))) | 23:04 |
@HeikoS | haha | 23:05 |
@HeikoS | generated-python- | 23:06 |
@wiking | okok | 23:06 |
@HeikoS | I guess that needs some clean up | 23:06 |
@HeikoS | but maybe later :) | 23:06 |
@wiking | r generated are currently not being run | 23:06 |
@wiking | right? | 23:06 |
@HeikoS | btw for the notebook github hook | 23:06 |
@HeikoS | that needs to be a new build on buildbot right? | 23:06 |
@wiking | just like lua | 23:06 |
@HeikoS | or can we us nightly default=? | 23:07 |
@wiking | right? | 23:07 |
@wiking | no | 23:07 |
@HeikoS | yeah | 23:07 |
@wiking | man wtf :) | 23:07 |
@wiking | i told you | 23:07 |
@HeikoS | lua and r are not ran | 23:07 |
@HeikoS | I forgot | 23:07 |
@wiking | check the PR hook build | 23:07 |
@wiking | http://buildbot.shogun-toolbox.org/builders/cookbook%20-%20PR | 23:07 |
@wiking | rename this | 23:07 |
@HeikoS | yeah | 23:08 |
@HeikoS | so in there | 23:08 |
@wiking | yeah you need to extend that job | 23:08 |
@wiking | is it generated-cpp | 23:08 |
@wiking | or generated_cpp | 23:08 |
@wiking | ? | 23:08 |
@wiking | and you should of course move it to maybe builder_pr or something like that | 23:12 |
@wiking | and just add a python modular build + the notebook steps | 23:12 |
@HeikoS | wiking: yeah I am trying to get through this stuff, it is a lot | 23:12 |
@wiking | thankyou | 23:12 |
@wiking | :) | 23:12 |
@HeikoS | generated_cpp-generated_cpp-generated_cpp- | 23:13 |
@HeikoS | ah | 23:13 |
@HeikoS | mouse | 23:13 |
@HeikoS | generated_cpp- | 23:13 |
@wiking | at least somebody knows how much energy was put into this shit | 23:13 |
@HeikoS | wiking: yeah :) | 23:13 |
@HeikoS | Ok i gotta go home now | 23:14 |
@HeikoS | Friday more | 23:14 |
@HeikoS | have to give a seminar tomorrow and havent prepared it yet | 23:14 |
@HeikoS | gnaaa | 23:14 |
@HeikoS | wiking: let me know when release is ready :) | 23:14 |
@HeikoS | ill check my mail | 23:14 |
@wiking | kk | 23:14 |
@wiking | i will release as soon as i can fix these little blockers | 23:14 |
@wiking | like why notebooks blocking the nightly upload | 23:14 |
@wiking | haha ok fuck | 23:15 |
@wiking | new thing | 23:15 |
@HeikoS | mmh something is weird | 23:15 |
@HeikoS | with the log | 23:15 |
@HeikoS | -P | 23:15 |
@HeikoS | option | 23:15 |
@HeikoS | ok I gotta go | 23:15 |
@HeikoS | good luck ;) | 23:15 |
@wiking | DEPECHEMODE | 23:16 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 23:16 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 23:16 | |
@wiking | shogun-buildbot: force build --branch=develop 'nightly_default' | 23:17 |
shogun-buildbot | build #13 forced | 23:17 |
shogun-buildbot | I'll give a shout when the build finishes | 23:17 |
@wiking | shogun-buildbot: force build --branch=develop 'bsd1 - libshogun' | 23:17 |
shogun-buildbot | build #39 forced | 23:17 |
shogun-buildbot | I'll give a shout when the build finishes | 23:17 |
shogun-buildbot | Hey! build bsd1 - libshogun #39 is complete: Success [build successful] | 23:18 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/bsd1%20-%20libshogun/builds/39 | 23:18 |
-!- OXPHOS [92bd305b@gateway/web/freenode/ip.146.189.48.91] has quit [Quit: Page closed] | 23:18 | |
-!- HeikoS [~heiko@untrust-out.swc.ucl.ac.uk] has quit [Ping timeout: 252 seconds] | 23:20 | |
shogun-buildbot | Hey! build nightly_default #13 is complete: Failure [failed notebooks] | 23:53 |
shogun-buildbot | Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/13 | 23:53 |
--- Log closed Thu Nov 03 00:00:59 2016 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!