[10:09:01] <sabya>	 is there a way to keep the SSH connection to labs instances alive for longer duration? I keep getting borken pipe after few mins.
[10:10:06] <sabya>	 ServerAliveInterval 3600 in .ssh/config did not help.
[16:22:17] <halfak>	 o/
[16:22:23] * halfak merges all of Amir1's work
[16:22:33] <Amir1>	 \o/
[16:22:34] <Amir1>	 \o/
[16:22:35] <Amir1>	 \o/
[16:22:37] <Amir1>	 \o/
[16:22:38] <Amir1>	 \o/
[16:22:42] <Amir1>	 I can go on
[16:23:09] <halfak>	 It's weird that we get fitness with SVC only after hours and hours of computation
[17:15:38] <halfak>	 Well... a moment of truth.  Do I kill the process that has been building SVC models for 50 hours and restart or let them go hoping they'll finish?
[17:19:51] <halfak>	 I went for the restart
[17:19:56] <halfak>	 We'll see how that goes. 
[17:38:18] <halfak>	 [[:en:Transformational leadership]]
[17:38:18] <AsimovBot>	 10[1] 04https://meta.wikimedia.org/wiki/:en:Transformational_leadership
[17:44:37] <travis-ci>	 wiki-ai/revscoring#579 (f1_test - 2bbdcd6 : halfak): The build passed. https://travis-ci.org/wiki-ai/revscoring/builds/115697996
[17:44:48] <halfak>	 Shuddup travis
[18:44:29] * halfak kicks off feature extraction for the new wikiclass models. 
[18:44:46] <halfak>	 We should have those ready for deployment in a couple of hours. 
[18:45:02] <halfak>	 There are only 4 (out of ~20) SVC models that aren't built yet. 
[18:45:10] <halfak>	 Hopefully my restart will be effective. 
[19:10:52] <halfak>	 oooh.  Testing the enwiki article quality model.  This will be interesting.  :) 
[19:11:00] <halfak>	 It's the first time I've seen f1 scores for each class. 
[19:17:56] <halfak>	 Looks like we're still riding the state-of-the-art.  c.f. http://arxiv.org/pdf/1603.01987v1.pdf
[19:55:10] <Amir1>	 halfak: ^
[19:55:15] <Amir1>	 :D