[MoE Refactor] Introduce RoutedExperts alias for FusedMoE and don't store SharedExperts in MK (#40735)

bnellnm committed 6 months ago

[Bugfix] [Frontend] Responses API, fix merging of messages (#42189)

yzong-rh committed 6 months ago

feat(kv-events): emit KV cache metadata (#40984)

Yan Ru Pei committed 6 months ago

[CPU] Fix rotary embedding for CPU without flash-attn ops (#42225)

Jonathan Mamou committed 6 months ago

[Refactor] Clean up pooling models `build_tok_params` logic (#42341)

Wentao Ye committed 6 months ago

[kv_offload][BugFix] Fix store deferral (#41945)

Martin Hickey committed 6 months ago

[docs] Added one new contact to the Vulnerability Management team (#42145)

Juan Pérez de Algaba committed 6 months ago

[vLLM IR] Minor improvements (#39362) (#39558)

Avishek Goswami committed 6 months ago

Added peagle speculators support (#41826)

shanjiaz committed 6 months ago

[Model] Support MiniCPM-V 4.6 (#41254)

tc-mb committed 6 months ago

[Build] Build bundled DeepGEMM `_C` per-Python so the wheel imports on every CPython (#41516)

Michael Goin committed 6 months ago

[Perf] Use 2D-grid to eliminate divmod in W8W8 group quant (#42153)

Jiahan Chang (Cyrus) committed 6 months ago

[Bugfix] Fix mismatched kernel-per-logical blocks in NIXL HMA transfer (#42097)

zhanqiuhu committed 6 months ago

[MoE Refactor] Move remaining experts classes to experts directory (#42334)

bnellnm committed 6 months ago

[MoE Refactor] Move expert map related code into ExpertMapManager class (#41046)

bnellnm committed 6 months ago

[Hybrid] Warmup Mamba2 SSD kernel (#39822)

Thomas Parnell committed 6 months ago

[MXFP4] Support for linear layers + compressed-tensors integration (#41664)

Dipika Sikka committed 6 months ago

[XPU] bump up vllm-xpu-kernels to v0.1.8 (#42410)

Kunshang Ji committed 6 months ago

[XPU] keep generator state of sycl kernel align with pytorch (#41771)

Yan Ma committed 6 months ago

[Doc] Fix typo in llm-d documentation link (#42397)

Florian Woerner committed 6 months ago