Skip to content

v0.4.0

Choose a tag to compare

@gujingit gujingit released this 23 Sep 03:49
· 319 commits to main since this release
Immutable release. Only release title and notes can be modified.
849757e

What's Changed

Features

  • feat: support rbgs scaling by @gujingit in #1
  • add workload status update event by @gujingit in #6
  • refactor: update dynamo demo; remove Chinese comments by @gujingit in #13
  • feat: Add pull request template by @bcfre in #11
  • add status check when diff workload by @gujingit in #15
  • feat: Format action templates to match sglang's pattern by @Pikabooboo in #19
  • add unit-tests by @gujingit in #28
  • support partition in rollingupdate by @gujingit in #30
  • feat: Add support for 1:1 rbg per topology assignment by @gujingit in #32
  • perf: reduce api-server load caused by exclusive-topology by @cheyang in #33
  • feature: support volcano podgroup by @ZYecho11 in #14

Bugfixs

  • bugfix: Fix the permission issue that rbgs controller cannot create rbg by @bcfre in #12
  • bugfix: Added consistency check for probes by @bcfre in #29

Build & CI

Docs

  • doc: Using SGLang as the default inference engine by @gujingit in #4
  • doc: Add CONTRIBUTING.md, development guide, updated image building logic by @bcfre in #9

New Contributors

Full Changelog: v0.3.0...v0.4.0