[06:44:17] Good morning and thank you all for the tips above. [07:05:53] good morning! [07:49:23] 06Machine-Learning-Team, 05Goal: Q4 24-25 Goal: Productionize peacock detection model - https://phabricator.wikimedia.org/T391940#10775375 (10isarantopoulos) [08:06:29] Good Morning. [08:07:05] klausman: Let me know if anything more info needed from our side for https://phabricator.wikimedia.org/T391958 and feel free to post questions/notes there as well. [08:12:46] ack will do. got a bit swamped yesterday [08:18:01] 10Lift-Wing, 06Machine-Learning-Team, 13Patch-For-Review: [onboarding] Improving language agnostic articlequality model + service - https://phabricator.wikimedia.org/T391679#10775471 (10OKarakaya-WMF) Performance test results from staging: ` (ozge) ozge@stat1010:~/inference-services/test/locust$ MODEL=arti... [08:49:32] klausman: Thanks! [08:59:16] (03PS1) 10Ozge: feat: adds locust results for articlequality_v2 [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1139790 [09:01:29] 10Lift-Wing, 06Machine-Learning-Team, 13Patch-For-Review: [onboarding] Improving language agnostic articlequality model + service - https://phabricator.wikimedia.org/T391679#10775567 (10OKarakaya-WMF) [09:09:15] 10Lift-Wing, 06Machine-Learning-Team, 13Patch-For-Review: [onboarding] Improving language agnostic articlequality model + service - https://phabricator.wikimedia.org/T391679#10775595 (10OKarakaya-WMF) I see the v2 latency is higher at 99.9% and retried it. I see we get similar results now. ` (ozge) ozge@s... [09:10:23] (03PS2) 10Ozge: feat: adds locust results for articlequality_v2 [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1139790 [09:11:46] One final patch to add performance test results above. I’ve shared the results in the task as well. https://phabricator.wikimedia.org/T391679#10775471 . In overall, we get similar results. Thank you all for your help during my onboarding and special thanks to my onboarding body @aiko :) [09:24:31] <3 [09:52:39] ozge_: thanks for all the great work done here! I have one last ask for the load tests : Since the models are served from the same pod it make sense to run a load test that includes both models. Do do that they would need to be in the same file and then we can define multiple users when running the test [09:53:12] wdyt? [09:53:29] if you want to see how to do this for multiple models it is similar to what we have in test/locust/models/revertrisk/revertrisk.py [10:07:38] Sure, thank you! I’m  updating it. [10:15:50] (03PS3) 10Ozge: feat: adds locust results for articlequality_v2 [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1139790 [10:18:42] (03PS4) 10Ozge: feat: adds locust results for articlequality_v2 [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1139790 [10:25:52] 10Lift-Wing, 06Machine-Learning-Team, 13Patch-For-Review: [onboarding] Improving language agnostic articlequality model + service - https://phabricator.wikimedia.org/T391679#10775818 (10OKarakaya-WMF) Running the models together: ` (ozge) ozge@stat1010:~/inference-services/test/locust$ MODEL=articlequality... [10:26:45] (03CR) 10Ilias Sarantopoulos: [C:03+1] "LGTM! Please add the Phabricator task in the commit msg as well" [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1139790 (owner: 10Ozge) [10:27:54] (03PS5) 10Ozge: feat: adds locust results for articlequality_v2 [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1139790 [10:28:50] (03PS6) 10Ozge: feat: adds locust results for articlequality_v2 [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1139790 [10:29:44] (03PS7) 10Ozge: feat: adds locust results for articlequality_v2 [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1139790 (https://phabricator.wikimedia.org/T391679) [10:31:28] @isaranto: added new results to the task and updated the patch. [10:34:32] (03CR) 10Ilias Sarantopoulos: [C:03+1] feat: adds locust results for articlequality_v2 [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1139790 (https://phabricator.wikimedia.org/T391679) (owner: 10Ozge) [10:37:51] (03CR) 10Ozge: [V:03+2 C:03+2] feat: adds locust results for articlequality_v2 [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1139790 (https://phabricator.wikimedia.org/T391679) (owner: 10Ozge) [11:09:03] * isaranto afk lunch [14:55:42] 06Machine-Learning-Team, 05Goal: Q4 24-25 Goal: Productionize peacock detection model - https://phabricator.wikimedia.org/T391940#10776860 (10isarantopoulos) We are currently trying to build an airflow pipeline for the data generation step as well as evaluation data for the human evaluation data so we can deci... [15:03:37] 10Lift-Wing, 06Machine-Learning-Team, 13Patch-For-Review: [onboarding] Improving language agnostic articlequality model + service - https://phabricator.wikimedia.org/T391679#10776894 (10OKarakaya-WMF) 05Open→03Resolved [15:13:37] hey folks, I filed https://gerrit.wikimedia.org/r/c/operations/deployment-charts/+/1139865 to upgrade knative on ml-serve-cofw [15:13:40] *codfw [15:13:45] lemme know if you are ok to test it [15:57:08] ack! [15:59:25] then we'd have to upgrade eqiad to bookworm and do the same thing right? [16:00:02] just asking to make sure I got this right [16:00:07] thanks [17:32:16] (03PS6) 10Sbisson: Support for articlecountry [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1137342 (https://phabricator.wikimedia.org/T391230)