[08:54:26] 10Lift-Wing, 06Machine-Learning-Team, 10Wikidata, 07OKR-Work: Optimize revertrisk-wikidata inference service to achieve ~500ms latency target - https://phabricator.wikimedia.org/T414060 (10kevinbazira) 03NEW [08:56:59] 06Machine-Learning-Team, 06Wikimedia Enterprise (WME Kanban): Test liftwing wikidata revert risk API for scale and latency - https://phabricator.wikimedia.org/T409388#11502440 (10kevinbazira) >>! In T409388#11499581, @FNavas-foundation wrote: > so sorry! yes sub-milisecond is incorrect. Half a second, 500ms.... [09:01:53] 06Machine-Learning-Team, 10Semantic Search: Semantic Search - Embeddings Service for MVP - https://phabricator.wikimedia.org/T412338#11502461 (10OKarakaya-WMF) [09:06:30] (03CR) 10Gkyziridis: [C:03+2] revert-risk: Treat 'testwiki' as 'enwiki' on the fly. [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1223639 (https://phabricator.wikimedia.org/T411786) (owner: 10Gkyziridis) [09:10:10] (03Merged) 10jenkins-bot: revert-risk: Treat 'testwiki' as 'enwiki' on the fly. [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1223639 (https://phabricator.wikimedia.org/T411786) (owner: 10Gkyziridis) [09:22:03] 06Machine-Learning-Team, 10Semantic Search: Semantic Search - Embeddings Service for MVP - https://phabricator.wikimedia.org/T412338#11502542 (10OKarakaya-WMF) Hello @dcausse , Do we plan to query the api on prod with the following prompt? We set max length to 300 chars. So that, if the query text length is h... [09:26:26] 06Machine-Learning-Team, 10Semantic Search: Semantic Search - Embeddings Service for MVP - https://phabricator.wikimedia.org/T412338#11502546 (10OKarakaya-WMF) We get better results on prod. - Median text length: 303. Requirement 72 (+ prompt if exists.) - Median latency 32ms. Max latency 290. Requirement <3... [09:54:43] 06Machine-Learning-Team, 10Semantic Search: Semantic Search - Embeddings Service for MVP - https://phabricator.wikimedia.org/T412338#11502601 (10dcausse) >>! In T412338#11502542, @OKarakaya-WMF wrote: > Hello @dcausse , > > Do we plan to query the api on prod with the following prompt? Yes, this is what is s... [10:12:25] 06Machine-Learning-Team, 10Semantic Search: Semantic Search - Embeddings Service for MVP - https://phabricator.wikimedia.org/T412338#11502650 (10OKarakaya-WMF) @dcausse , cool. I'll update the service and the performance tests accordingly. [11:02:03] (03PS1) 10Ozge: feat: embeddings server performance tests [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1224606 [11:05:16] (03PS2) 10Ozge: feat: embeddings server performance tests [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1224606 [11:13:47] (03CR) 10Kevin Bazira: [C:03+1] feat: embeddings server performance tests [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1224606 (owner: 10Ozge) [11:17:32] (03CR) 10Ozge: [C:03+2] feat: embeddings server performance tests [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1224606 (owner: 10Ozge) [11:17:50] (03CR) 10Ozge: [V:03+2 C:03+2] feat: embeddings server performance tests [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1224606 (owner: 10Ozge) [11:18:55] (03Merged) 10jenkins-bot: feat: embeddings server performance tests [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1224606 (owner: 10Ozge) [13:05:28] 06Machine-Learning-Team, 10Semantic Search: Semantic Search - Embeddings Service for MVP - https://phabricator.wikimedia.org/T412338#11503278 (10OKarakaya-WMF) when median length of the query: 77 chars + the prompt (108 chars): - max latency: 290ms. - 99.9 percentile latency: 280ms. - median latency: 34ms... [13:21:52] 06Machine-Learning-Team, 06Data-Platform-SRE (2026.01.05 - 2026.01.23), 06Editing-team (Tracking), 07Essential-Work: Build model training pipeline for tone check using WMF ML Airflow instance - https://phabricator.wikimedia.org/T396495#11503510 (10Gehel) [14:07:38] 06Machine-Learning-Team, 10Semantic Search: Semantic Search - Embeddings Service for MVP - https://phabricator.wikimedia.org/T412338#11503757 (10dcausse) @OKarakaya-WMF awesome thanks, p50 at 34ms is nice thanks! If I'm reading the numbers right it does seem like the prompt is not adding much overhead. yes the... [15:24:33] 06Machine-Learning-Team, 10Semantic Search: Semantic Search - Embeddings Service for MVP - https://phabricator.wikimedia.org/T412338#11504098 (10OKarakaya-WMF) I agree! thank you @dcausse [16:18:50] (03PS1) 10Sbisson: Guard against: 'NoneType' object has no attribute 'keys' [research/recommendation-api] - 10https://gerrit.wikimedia.org/r/1224716 [16:33:16] 06Machine-Learning-Team, 06Data-Persistence, 10Data-Persistence-Design-Review, 06Growth-Team, and 3 others: Data Persistence Design Review: Improve Tone Suggested Edits newcomer task - https://phabricator.wikimedia.org/T401021#11504452 (10Ahoelzl) [16:39:36] 06Machine-Learning-Team, 05Goal, 07OKR-Work: Q1 FY2025-26 Goal: Task generation engine for Revise Tone task - https://phabricator.wikimedia.org/T408341#11504540 (10achou) [16:39:39] 06Machine-Learning-Team, 06Growth-Team, 10Revise-Tone-Structured-Task, 13Patch-For-Review: Use HTML instead of wikitext for Revise Tone Task Generator in LiftWing - https://phabricator.wikimedia.org/T412210#11504539 (10achou) [17:38:19] 06Machine-Learning-Team, 06Wikimedia Enterprise (WME Kanban): Test liftwing wikidata revert risk API for scale and latency - https://phabricator.wikimedia.org/T409388#11504749 (10FNavas-foundation) @kevinbazira thank you. [21:11:08] (03PS1) 10Kgraessle: When filtering for edits with high Revert Risk, Recent Changes shouldn't display edits from non-main namespaces [extensions/ORES] - 10https://gerrit.wikimedia.org/r/1224813 (https://phabricator.wikimedia.org/T404200) [22:27:33] 06Machine-Learning-Team, 06Data-Persistence, 10Data-Persistence-Design-Review: Data Persistence Design Review: Article topic model caching - https://phabricator.wikimedia.org/T402984#11505672 (10Eevans)