[02:08:38] (03CR) 10Sbisson: Support Default collections (032 comments) [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1072175 (https://phabricator.wikimedia.org/T374597) (owner: 10Santhosh) [06:50:08] 06Machine-Learning-Team: [ml-lab] Use a (jupyter) notebook and load a LLM from huggingface - https://phabricator.wikimedia.org/T377574#10291566 (10isarantopoulos) Given a virtual environment that is created under the user's home dir with: `python3 -m venv .venv` and after installing the following requirements wi... [08:42:34] (03PS4) 10Kevin Bazira: test: automate running unit tests in the LW repo [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) [08:43:09] (03CR) 10Kevin Bazira: "Following our discussion on IRC, this patch has been updated to run tests within the blubber test variant of a model-server starting with " [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) (owner: 10Kevin Bazira) [08:47:37] isaranto: o/ [08:47:37] I have implemented 2.1 from our discussion yesterday. have a look whenever you get a minute: [08:47:37] https://gerrit.wikimedia.org/r/c/machinelearning/liftwing/inference-services/+/1087146/comments/b1a52242_837aa900 [08:47:49] o/ kevin [08:48:45] good morning, that's great, thanks! I will be able to review a bit later [08:49:37] quick question: do we need to suppress the warnings? what I want to say is that if ci passes and no error is thrown then we should better leave them there [08:50:18] otherwise if the exit code is 1 I get it [08:55:38] morning morning! [08:55:39] the warnings were being suppressed because they are being thrown by third-party dependencies like pytools [08:56:26] would you prefer we don't suppress them? [09:16:33] 06Machine-Learning-Team, 06Data-Engineering, 06Research, 10Event-Platform: Expose revision revert risk scores in EventStreams - https://phabricator.wikimedia.org/T326179#10291750 (10kostajh) > 1. Create a new mediawiki.page_revert_risk_prediction_change.v1 stream, and produce events to it, either via chang... [09:46:35] morning! [09:58:27] Morning Aiko! [09:59:52] kevinbazira: yes, in general I think it is good to show the warnings. Sometimes they might be worth fixing or inform us on future deprecation plans etc [10:09:26] (03PS5) 10Kevin Bazira: test: automate running unit tests in the LW repo [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) [10:11:07] okok I've removed the filter that suppresses warnings --^ [10:15:46] 🙌 [10:24:41] 06Machine-Learning-Team: Test the feasibility of deployment of Aya-23 model in LiftWing - https://phabricator.wikimedia.org/T379052 (10MGerlach) 03NEW [10:36:07] 06Machine-Learning-Team: Test the feasibility of deployment of Aya-23 model in LiftWing - https://phabricator.wikimedia.org/T379052#10291972 (10MGerlach) Update: [[ https://huggingface.co/CohereForAI/aya-23-8B | Aya-23-8B model]] runs successfully in LiftWing. thanks @isarantopoulos Example query (works only in... [10:40:44] (03CR) 10Ilias Sarantopoulos: "Great work Kevin! I added some comments so that we can work in the direction to also make sure CI will be able to run for the rest of the " [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) (owner: 10Kevin Bazira) [10:45:59] 06Machine-Learning-Team: Test the feasibility of deployment of Aya-23 model in LiftWing - https://phabricator.wikimedia.org/T379052#10292028 (10MGerlach) Update: @isarantopoulos did first experiments with the Aya-23-35B model. It does not work out of the box. The raw version of the model is 65GB on disk and does... [11:12:34] (03PS6) 10Kevin Bazira: test: automate running unit tests in the LW repo [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) [11:14:22] (03CR) 10Kevin Bazira: test: automate running unit tests in the LW repo (032 comments) [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) (owner: 10Kevin Bazira) [11:25:07] * isaranto afk lunch [12:30:36] (03PS1) 10AikoChou: reference-quality: update output schema [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087446 (https://phabricator.wikimedia.org/T378939) [12:43:32] (03CR) 10Nik Gkountas: Support Default collections (032 comments) [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1072175 (https://phabricator.wikimedia.org/T374597) (owner: 10Santhosh) [12:43:47] (03PS23) 10Nik Gkountas: Support Default collections [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1072175 (https://phabricator.wikimedia.org/T374597) (owner: 10Santhosh) [13:31:32] (03CR) 10Ilias Sarantopoulos: test: automate running unit tests in the LW repo (032 comments) [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) (owner: 10Kevin Bazira) [13:59:19] (03PS7) 10Kevin Bazira: test: automate running unit tests in the LW repo [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) [14:02:18] (03CR) 10Kevin Bazira: test: automate running unit tests in the LW repo (031 comment) [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) (owner: 10Kevin Bazira) [15:21:07] 06Machine-Learning-Team: ml-lab should have documentation - https://phabricator.wikimedia.org/T376974#10292948 (10klausman) https://wikitech.wikimedia.org/wiki/Machine_Learning/ML-Lab has been created with some basic information. [15:23:04] 06Machine-Learning-Team, 13Patch-For-Review: Update output schema for reference risk model - https://phabricator.wikimedia.org/T378939#10292965 (10isarantopoulos) a:03achou [15:25:42] 06Machine-Learning-Team, 13Patch-For-Review: Update output schema for reference risk model - https://phabricator.wikimedia.org/T378939#10292967 (10achou) [15:26:12] 06Machine-Learning-Team: Possibility to transition ml-serve[2001-2008] and and ml-staging[2001-2002] to 10G - https://phabricator.wikimedia.org/T378714#10292969 (10isarantopoulos) [15:26:50] 06Machine-Learning-Team: Possibility to transition ml-serve[2001-2008] and and ml-staging[2001-2002] to 10G - https://phabricator.wikimedia.org/T378714#10292970 (10klausman) a:03klausman [15:56:08] (03PS8) 10Kevin Bazira: test: automate running unit tests in the LW repo [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) [15:58:30] (03CR) 10Kevin Bazira: test: automate running unit tests in the LW repo (031 comment) [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) (owner: 10Kevin Bazira) [15:59:33] (03CR) 10Sbisson: [C:03+2] Support Default collections [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1072175 (https://phabricator.wikimedia.org/T374597) (owner: 10Santhosh) [16:00:15] (03Merged) 10jenkins-bot: Support Default collections [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1072175 (https://phabricator.wikimedia.org/T374597) (owner: 10Santhosh) [16:30:02] 06Machine-Learning-Team, 06Data-Engineering, 06Research, 10Event-Platform: Expose revision revert risk scores in EventStreams - https://phabricator.wikimedia.org/T326179#10293306 (10Ottomata) Hm, yes I think both could be done! However, where should the 'fattening' of recentchange event data stop? Shou... [17:12:30] 06Machine-Learning-Team, 10Observability-Metrics, 10SRE Observability (FY2024/2025-Q2): Gap in metrics rendered from Thanos Rules - https://phabricator.wikimedia.org/T352756#10293599 (10lmata) [17:36:36] going afk folks! cu tomorrow [17:40:18] (03CR) 10Ilias Sarantopoulos: [C:03+1] "Nice, great work ! 🎉 this would seem to work. Please build one of the other test images + change the commit message before merging. thanks" [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) (owner: 10Kevin Bazira) [19:23:11] 06Machine-Learning-Team, 13Patch-For-Review: Run unit tests for the inference-services repo in CI - https://phabricator.wikimedia.org/T360120#10294102 (10kevinbazira) Running unit tests has been automated. When a model server's test variant has all the necessary files (model-server, python utils, unit tests, t... [19:39:14] (03PS1) 10Nik Gkountas: Add articles_count to page collection metadata [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1087559 [19:54:26] (03CR) 10Eamedina: [C:03+2] Add articles_count to page collection metadata [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1087559 (owner: 10Nik Gkountas) [19:55:06] (03Merged) 10jenkins-bot: Add articles_count to page collection metadata [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1087559 (owner: 10Nik Gkountas) [20:11:48] (03PS9) 10Kevin Bazira: test: automate running unit tests in the LW repo [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1087146 (https://phabricator.wikimedia.org/T360120) [20:47:53] (03PS1) 10Nik Gkountas: Add articles_by_language_count field to metadata [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1087569