pullrequests Search Results · language:Dune language:Python language:Java language:Java language:Java language:Java language:Java
Filter by
88.6M results
Automated release preparation for v0.7.3.
Merging this PR creates the v0.7.3 tag at the merge commit and publishes the GitHub Release plus the signed multi-arch
images. Closing it without merging cancels ...
release
In rare cases the Triggerer subprocess can lose track of a trigger — it is assigned in the supervisor s running_triggers
but no coroutine is actually running. The affected task instance stays in deferred ...
area:Triggerer
概要
配套 runwhere-ai 控制台(runwhere-ai/runwhere-ai#2)的 gpuctl 侧改动,合入 PR #6 之后累积的两笔。
内容
任务遥测 sidecar(5af3ea7)
- GPU 任务 Pod 注入 opt-in 原生 sidecar(由 GPUCTL_TELEMETRY_ENDPOINT feature-flag 控制),仅在容器申请 GPU 时注入;无 ...
What
The LLM-as-judge functions silently fell back to a 0.5 score whenever DeepSeek wrapped its JSON reply in a Markdown code
fence (json … ). A bare json.loads() raised JSONDecodeError, the except returned ...
Changelog Description
- Change the title of MK Doc to # AYON Deadline Addon API Reference
Testing notes:
I triggered the action from cli via gh workflow run deploy_mkdocs.yml --ref chore/258-change-title-of-mk-docs ...
type: maintenance
What
Adjudicates the three dependency-free blocker-resolution lane reports (DU-1 raw admission, DU-2 write-boundary, DU-6
actor governance) and files the results as sibling artifacts, closing the gating ...
Summary
- Add read-only adaptive-rag chat sessions list and adaptive-rag chat sessions show commands.
- Reuse the M14 chat history API schemas for stable JSON output from the CLI.
- Preserve project-scoped ...
What
Fixed vLLM backend failing with LLM.generate() is only supported for generative models error when evaluating base
generative models (e.g. norallm/norbloom-7b-scratch). The vLLM LLM constructor now ...
What why
Importing simkl_mps ran simkl_mps/__init__.py:13 → from simkl_mps.tray_win import run_tray_app, which unconditionally
imported the Windows tray module on every platform. That chain (tray_win ...
feature
size/s
workflow