[06:59:34] <isaranto>	 Morning!
[07:38:20] <isaranto>	 unfortunately the patch to be able to edit revisions and isvcs in experimental ns doesnt seem to work, I get "forbidden"
[08:59:35] <aiko>	 morning Ilias o/
[09:02:53] <isaranto>	 Hey aiko o/
[12:09:34] * isaranto afk lunch
[12:13:13] <wikibugs>	 10Machine-Learning-Team: Support running revertrisk-multilingual model-server via Makefile - https://phabricator.wikimedia.org/T356501 (10achou)
[12:16:09] <wikibugs>	 (03PS1) 10AikoChou: Makefile: add support for revertrisk-multilingual [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/995198 (https://phabricator.wikimedia.org/T356501)
[14:21:32] <chrisalbon>	 Good morning
[14:26:11] <isaranto>	 Buongiorno!
[14:30:53] <isaranto>	 aiko: ping me if you need a review for the above --^ at the moment I tried to run it but got an error
[14:35:20] <aiko>	 isaranto: ah what error did you get?
[14:35:46] <isaranto>	 https://phabricator.wikimedia.org/P56124
[14:35:52] <isaranto>	 I ran `make run`
[14:36:21] <wikibugs>	 (03PS1) 10AikoChou: revertrisk: use GPU for revertrisk-multilingual [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/995214 (https://phabricator.wikimedia.org/T356045)
[14:36:22] <isaranto>	 I didnt do much just checked out the branch and tested it. lemme know if you need help and/or if I should deep dive a bit more
[14:43:22] <wikibugs>	 (03CR) 10CI reject: [V: 04-1] revertrisk: use GPU for revertrisk-multilingual [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/995214 (https://phabricator.wikimedia.org/T356045) (owner: 10AikoChou)
[14:48:09] <aiko>	 I had that error before, and then I switched to py 3.10 and the problem  disappeared 
[14:50:05] <wikibugs>	 10Lift-Wing, 10Machine-Learning-Team: Debug GPU deployments on ml-staging - https://phabricator.wikimedia.org/T356038 (10isarantopoulos) For the moment I came up with the following 2 options to implement the above functionality:  - Enable hot reload for the inference service: since kserve is based on FastAPI w...
[14:51:34] <isaranto>	  but we use python 3.9 ¯\_(ツ)_/¯
[14:53:24] <wikibugs>	 (03PS2) 10AikoChou: revertrisk: use GPU for revertrisk-multilingual [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/995214 (https://phabricator.wikimedia.org/T356045)
[14:54:03] <aiko>	 ahh let me try py 3.9 
[15:02:31] <isaranto>	 I added my thoughts for GPU testing on the task above.lemme know your thoughts. aiko: if you want to wait for me to try this properly on Monday and then we can also try rrml model this way
[15:03:41] <isaranto>	 actually nevermind your tests are not related, we can proceed testing rrml separately anyway. The difference is that if sth fails this way we could iterate faster
[15:19:20] <aiko>	 testing rrml is not urgent. I'll wait for you trying things out. let me know if I can help with anything on monday!
[15:23:32] <aiko>	 and I'm not sure if the patch I made will work or fail, so it'd be nice if we can fast iterate
[15:24:17] <isaranto>	 cool cool cool cool
[15:24:21] <isaranto>	 :)
[15:35:40] <elukey>	 hey folks, I am reading https://nostarch.com/how-ai-works, it is really really nice 
[15:40:21] <isaranto>	 Hey Luca! Interested to discuss it or even read it after you're done!
[15:43:19] <isaranto>	 Logging off for the weekend folks, I stumbled in some issues with the article description load tests so I'll report results on Monday morning ( trying to set up a test file  with different hosts)
[16:01:53] <aiko>	 hey Luca, thanks for sharing :))
[16:02:12] <aiko>	 o/ have a nice weekend folks
[17:15:58] <chrisalbon>	 elukey!
[17:16:06] <chrisalbon>	 have a great weekend all