[04:47:12] 10Lift-Wing, 06Machine-Learning-Team, 07Essential-Work: [LLM] Lift Wing load testing - https://phabricator.wikimedia.org/T377225#12288277 (10isarantopoulos) 05Open→03Declined This is an old task which I'm going to decline. A lot has happened since then. For benchmarking or load testing LLM services w... [04:50:11] 10Lift-Wing, 06Machine-Learning-Team (Q1 FY2026-27), 13Patch-For-Review: Host Qwen3.8-27B-FP8 on LiftWing - https://phabricator.wikimedia.org/T436644#12288281 (10isarantopoulos) Great work! Instead of this custom script could you run vllm-bench like we did for the other models in https://wikitech.wikimedia.... [04:56:09] 10Lift-Wing, 06Machine-Learning-Team (Q1 FY2026-27), 13Patch-For-Review: Host Qwen3.8-27B-FP8 on LiftWing - https://phabricator.wikimedia.org/T436644#12288286 (10isarantopoulos) Seems like we have the whole process documented in the readme markdown file in https://gerrit.wikimedia.org/r/plugins/gitiles/machi... [05:13:02] 06Machine-Learning-Team (Q1 FY2026-27), 06Data-Platform-SRE (2026-08-28 - 2026-09-18): monitoring: Grafana dashboard for LLM serving on MI300X - https://phabricator.wikimedia.org/T429237#12288302 (10isarantopoulos) I see that there are already 2 user dashboards created: @DPogorzelski-WMF has created this one... [05:13:48] 06Machine-Learning-Team (Q1 FY2026-27): [draft] Q1 FY2026-27 LLM Platform Milestone 2 - https://phabricator.wikimedia.org/T433975#12288305 (10isarantopoulos) [05:13:49] 06Machine-Learning-Team (Q1 FY2026-27), 06Data-Platform-SRE (2026-08-28 - 2026-09-18): monitoring: Grafana dashboard for LLM serving on MI300X - https://phabricator.wikimedia.org/T429237#12288304 (10isarantopoulos) [05:13:50] 06Machine-Learning-Team (Q1 FY2026-27): Q1 FY2026-27 Milestone 1: Deliver a V0 LLM platform for Wikimania - https://phabricator.wikimedia.org/T431846#12288306 (10isarantopoulos) [05:14:28] 06Machine-Learning-Team (Q1 FY2026-27): [draft] Q1 FY2026-27 LLM Platform Milestone 2 - https://phabricator.wikimedia.org/T433975#12288308 (10isarantopoulos) [05:14:29] 06Machine-Learning-Team (Q1 FY2026-27): Q1 FY2026-27 Milestone 1: Deliver a V0 LLM platform for Wikimania - https://phabricator.wikimedia.org/T431846#12288309 (10isarantopoulos) [05:14:30] 06Machine-Learning-Team (Q1 FY2026-27), 06Data-Platform-SRE (2026-08-28 - 2026-09-18): monitoring: View GPU usage per LLM deployment/model - https://phabricator.wikimedia.org/T429236#12288307 (10isarantopoulos) [05:15:49] 06Machine-Learning-Team (Q1 FY2026-27): Q1 FY2026-27 Milestone 1: Deliver a V0 LLM platform for Wikimania - https://phabricator.wikimedia.org/T431846#12288311 (10isarantopoulos) 05Open→03Resolved a:03isarantopoulos Congrats everyone for all the work done here! Resolving this and moving 2 unresolved sub... [05:21:39] 06Machine-Learning-Team (Q1 FY2026-27), 07Essential-Work: liftwing-openapi-server: Add the LLM endpoints to the OpenAPI spec - https://phabricator.wikimedia.org/T434682#12288318 (10isarantopoulos) 05Open→03Resolved [06:30:03] 06Machine-Learning-Team (Q1 FY2026-27), 06Data-Platform-SRE (2026-08-28 - 2026-09-18): monitoring: Grafana dashboard for LLM serving on MI300X - https://phabricator.wikimedia.org/T429237#12288390 (10kevinbazira) >>! In T429237#12288301, @isarantopoulos wrote: > I see that there are already 2 user dashboards cr... [07:53:11] (03CR) 10Bartosz Wójtowicz: "Thank you for the work here <3 Leaving a couple of comments" [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1334976 (https://phabricator.wikimedia.org/T436892) (owner: 10Kevin Bazira) [09:05:57] (03CR) 10Ozge: Add semantic_highlighting (032 comments) [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1322788 (https://phabricator.wikimedia.org/T433872) (owner: 10DCausse) [09:22:59] (03CR) 10Ilias Sarantopoulos: "Thanks for working on this Kevin!" [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1334976 (https://phabricator.wikimedia.org/T436892) (owner: 10Kevin Bazira) [09:29:56] (03CR) 10Ilias Sarantopoulos: tool_server: add tool-server exposing REST and MCP (031 comment) [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1334976 (https://phabricator.wikimedia.org/T436892) (owner: 10Kevin Bazira) [10:25:57] (03CR) 10Ozge: Add semantic_highlighting (034 comments) [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1322788 (https://phabricator.wikimedia.org/T433872) (owner: 10DCausse) [12:39:09] 10Lift-Wing, 06Machine-Learning-Team (Q1 FY2026-27): Benchmark Qwen3.8-27B-FP8 model - https://phabricator.wikimedia.org/T437034 (10gkyziridis) 03NEW [12:48:03] (03PS1) 10Gkyziridis: benchmarks: add llm-qwen38-27b baseline [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1335839 (https://phabricator.wikimedia.org/T437034) [12:49:43] (03CR) 10CI reject: [V:04-1] benchmarks: add llm-qwen38-27b baseline [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1335839 (https://phabricator.wikimedia.org/T437034) (owner: 10Gkyziridis) [12:52:37] 10Lift-Wing, 06Machine-Learning-Team (Q1 FY2026-27), 13Patch-For-Review: Host Qwen3.8-27B-FP8 on LiftWing - https://phabricator.wikimedia.org/T436644#12289530 (10gkyziridis) I created a subtask for the benchmark: https://phabricator.wikimedia.org/T437034 adding my findings there. [12:53:07] 10Lift-Wing, 06Machine-Learning-Team (Q1 FY2026-27), 13Patch-For-Review: Benchmark Qwen3.8-27B-FP8 model - https://phabricator.wikimedia.org/T437034#12289534 (10gkyziridis) The vllm-bench results can be found in the following paste: {P96348} [12:56:50] (03PS2) 10Gkyziridis: benchmarks: add llm-qwen38-27b baseline [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1335839 (https://phabricator.wikimedia.org/T437034) [13:02:27] 06Machine-Learning-Team, 10VisualEditor, 13Patch-For-Review: Integrating real time edit checks in Visual Editor using ML - https://phabricator.wikimedia.org/T393301#12289567 (10PatchDemoBot) Test wiki on [[ https://patchdemo.wmcloud.org/ | Patch demo ]] by ISarantopoulos-WMF using patch(es) linked to this t... [13:05:36] 06Machine-Learning-Team, 10VisualEditor, 13Patch-For-Review: Integrating real time edit checks in Visual Editor using ML - https://phabricator.wikimedia.org/T393301#12289607 (10PatchDemoBot) Test wiki on [[ https://patchdemo.wmcloud.org/ | Patch demo ]] by ISarantopoulos-WMF using patch(es) linked to this t... [13:07:34] 06Machine-Learning-Team, 10VisualEditor, 13Patch-For-Review: Integrating real time edit checks in Visual Editor using ML - https://phabricator.wikimedia.org/T393301#12289646 (10PatchDemoBot) Test wiki on [[ https://patchdemo.wmcloud.org/ | Patch demo ]] by ISarantopoulos-WMF using patch(es) linked to this t... [15:53:21] (03PS2) 10Kevin Bazira: tool_server: add tool-server exposing REST and MCP [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1334976 (https://phabricator.wikimedia.org/T436892) [15:56:52] (03CR) 10Kevin Bazira: "I think these a great suggestions, thanks! I've added `errors.py` and now `service.py` discovers tools in their directories under `tools/`" [machinelearning/liftwing/inference-services] - 10https://gerrit.wikimedia.org/r/1334976 (https://phabricator.wikimedia.org/T436892) (owner: 10Kevin Bazira)