v0.55.0
What's Changed
- feature: add support to skip jobs that don't have modified files in CI by @ndizazzo in #116
- launch: asymmetric KV cache — Q8_0 K + Q4_0 V for 5-50GB models by @michaelneale in #104
- Fix compilation errors from missing
served_model_descriptorsin test struct initializers by @Copilot in #120 - Fix compile errors in model descriptor test helpers by @Copilot in #121
- Fix UI build to fail on TypeScript compile errors by @i386 in #124
- Add runtime model load/unload with CLI and REST by @i386 in #56
- URGENT: Fix client --auto join hang (connect_to_peer regression) by @michaelneale in #126
- ci: add client-auto boot test by @michaelneale in #125
- Add canonical served model descriptors by @i386 in #118
- Decomposing the codebase by @i386 in #127
- Remove Windows builds from CI by @i386 in #140
- router: context-aware host selection and retry on context overflow by @michaelneale in #141
Full Changelog: v0.54.0...v0.55.0