[KV Connector][Offloading] Keep per-layer KV registration when canonical_layout is requested (#51688)

Itay Etelis committed 3 months ago

[Bugfix][ROCm][CI] Stabilize build context and source caches (#51721)

Andreas Karatzas committed 3 months ago

[Bugfix][Model Loader] Defer post-load attention weight processing (#49519)

aoshen02 committed 3 months ago

[Bugfix] Import each packed IPC export once on the consumer side (#51259)

acmore committed 3 months ago

[Frontend] Add content_parts to /inference/v1/generate for raw multim… (#51478)

aoshen02 committed 3 months ago

[CI] Solidify speculative decoding E2E coverage (#50713)

Andreas Karatzas committed 3 months ago

Fix chat completion 500 on non-object JSON bodies (#51654)

Tarun Kumar committed 3 months ago

[Rust Frontend] Support dynamic tools from developer messages (#51144)

Bugen Zhao committed 3 months ago

[MM][CG][BugFix] Fix Ernie-4.5-VL encoder CG postprocess for multi-path outputs (#51461)

Qiuyang Yue committed 3 months ago

[Rust Frontend] Upgrade MiniJinja to 2.22 & remove method lookup workaround (#51235)

Bugen Zhao committed 3 months ago

[CI] Parallelize release image publishing (#51735)

Kevin H. Luu committed 3 months ago

[Perf] Skip detokenization in offline beam search (#50333)

Myeongwon Kim committed 3 months ago

[Bugfix] Fix DeepSeek V4/3.2 tokenizer vocab size overcount crashing guided decoding (#51727)

Flora Feng committed 3 months ago

[Model Runner V2][MTP] Share topk index buffer between draft steps (#47352)

Giancarlo Delfin committed 3 months ago

[Bugfix][MiMo] Apply vision attention sinks in the window attention path (#49815)

Almog Tavor committed 3 months ago

[Bugfix] Preserve non-logitproc entry points in tests (#51097)

Tristan Rice committed 3 months ago

[Rust Frontend][gRPC] Add explicit data-parallel rank routing (#51178)

Connor Carpenter committed 3 months ago

[CI] Add /ci cancel command (#51732)

Kevin H. Luu committed 3 months ago

[Build][gRPC] Publish protobuf schemas to Buf (#51276)

Connor Carpenter committed 3 months ago

[Kimi-K3] DCP support (#50484)

Summer Yang committed 3 months ago