[08:06:24] <travis-ci>	 wiki-ai/revscoring#1334 (multilabel-rf - 817d6f6 : Sumit Asthana): The build was fixed. https://travis-ci.org/wiki-ai/revscoring/builds/314237721
[14:10:58] <Zppix>	 halfak: awight gooood mornijg
[14:11:03] <Zppix>	 Morning*
[14:24:46] <cyberjedi>	 Was i here asking if i could deploy an AI Engine here?? been to so many
[14:25:06] <cyberjedi>	 room caught my eye
[14:26:00] <Zppix>	 cyberjedi: you may not
[14:26:07] <Zppix>	 cyberjedi: please leave
[14:26:42] <Zppix>	 Hes a troll
[14:26:52] <Zppix>	 Hes been in wikipedia-en before
[14:29:32] <refeed[m]>	 I've seen him at #wikidata also :/
[15:40:44] <travis-ci>	 wiki-ai/revscoring#1336 (multilabel-rf - 9f88d40 : Sumit Asthana): The build was broken. https://travis-ci.org/wiki-ai/revscoring/builds/314357828
[15:44:49] <travis-ci>	 wiki-ai/revscoring#1338 (setuppy-fix - d8b1ce0 : Sumit Asthana): The build failed. https://travis-ci.org/wiki-ai/revscoring/builds/314359039
[15:50:21] <travis-ci>	 wiki-ai/revscoring#1341 (multilabel-rf - be9d151 : Sumit Asthana): The build was fixed. https://travis-ci.org/wiki-ai/revscoring/builds/314360453
[16:22:12] <Zppix>	 codezee: i commented on your PR for revscoring
[17:27:19] <codezee>	 halfak: o/
[17:36:11] <halfak>	 o/ codezee 
[17:36:36] <halfak>	 What do you think of my addition to your PR
[17:36:37] <halfak>	 ?
[17:36:51] <halfak>	 Ahh I see you have more commits :) 
[17:37:47] <halfak>	 codezee, I don't think it's a good idea to not pass in the labels during initialization. 
[17:37:59] <halfak>	 That was the primary reason I moved away from the sklearn binarizer. 
[17:38:28] <halfak>	 Also, "check label consistency" means nothing in that case. 
[17:39:06] <Zppix>	 Halfak ill review codezees revs pr since it touches what i did with py3 requirements okay?
[17:39:23] <halfak>	 Zppix, if you want to leave some notes, that's fine. 
[17:39:36] <halfak>	 Oh you're talking about a different PR
[17:39:50] <Zppix>	 Yes :)
[17:39:51] <Zppix>	 Revs
[17:40:09] <Zppix>	 Its a non trival
[17:40:14] <codezee>	 halfak: yeah wanted to talk about that
[17:40:35] <halfak>	 codezee, "labels" is a required param of classification Models. 
[17:41:17] <codezee>	 halfak: that was the issue i faced with postponed label initialization
[17:41:26] <Zppix>	 codezee: merged ty
[17:41:33] <codezee>	 Zppix: :_
[17:41:37] <codezee>	 * :)
[17:42:35] <codezee>	 halfak: but doesn't it makes sense to initialize labels during fitting otherwise we're writing additional code  in tune to infer labels?
[17:43:31] <codezee>	 halfak: also i see that sklearns binarizer can take classes as input  - https://github.com/scikit-learn/scikit-learn/blob/a24c8b46/sklearn/preprocessing/label.py#L678
[17:43:38] <halfak>	 Oh we provide labels in tune
[17:43:39] <halfak>	 We must 
[17:43:46] <halfak>	 It's a required param of classification models. 
[17:44:06] <halfak>	 codezee, gotcha.  That was undocumented. 
[17:44:32] <halfak>	 Oh nevermind.  I see it in the docs now. 
[17:44:45] <halfak>	 Regardless, the label normalizer is a nice and useful abstraction :) 
[17:45:32] <codezee>	 halfak: yeah abstraction part is true
[17:47:52] <codezee>	 halfak: does it make sense to have sklearn's binarizer utility in the custom binarizer - that of sklearn uses numpy functions so i'm thinking performance might be a thing 
[17:48:49] <halfak>	 codezee, I can't imagine it would be a big thing since I'm using efficient data structures, but if you want to run a test, I'd be interested in the result. 
[17:49:36] <codezee>	 halfak: ok, i think it'll be best to create a task and keep it in queue, right now lets get this in and test
[17:49:55] <codezee>	 i'll edit the pr to restore the initialization thing
[17:50:06] <halfak>	 cool.  Then I think we're good to go. 
[17:52:23] <halfak>	 codezee, if you restore that and the tests are passing in travis, feel free to self-merge.  I'm going to step away for a bit, but I don't want to delay you :) 
[17:52:39] <halfak>	 Anything you need from me before I run away. 
[17:52:44] <Zppix>	 (I can merge too)
[17:54:52] <codezee>	 i don't think i need anything rn, got it covered
[17:54:57] <Zppix>	 Ok
[17:55:09] <halfak>	 Zppix, this one's pretty big.  I think it's better if we have you review/merge PRs that are with bits you have experience with :) 
[17:55:21] <Zppix>	 Ok
[17:56:59] <halfak>	 OK I'm out of here.  Have a good one folks! o/
[18:02:28] <Zppix>	 You too