[1/N][KV Connector] Identify externally transferable KV cache groups (#53779)

Matthew Bonanni committed about 2 months ago

buffer size insuffient Dspark sd for FlashInfer MNNVL allreduce (#50932)

Khushali Desai committed about 2 months ago

[Elastic EP] Preserve AOT cache reuse during scaling (#53378)

Itay Alroy committed about 2 months ago

[Bugfix] Update FlashMLA for sparse decode workspace fix (#53755)

Matthew Bonanni committed about 2 months ago

[Bugfix][Frontend] Let pooling requests set padding (#51157)

Hert4 committed about 2 months ago

fix(config): guard LlamaBidirectionalConfig against missing hf_config.pooling (#50536)

hcl committed about 2 months ago

[Bugfix][KV Offload] Handle padded GPU cache storage (#54021)

Jiangyun Zhu committed about 2 months ago

[RL] Support checkpoint-coordinate sparse NCCL weight updates (#53751)

Shuolei Wang committed about 2 months ago

[Bugfix][OpenAI] Fix streamed completion logprob offsets with echo (#47815)

Ting SUN committed about 2 months ago

[Bugfix] Raise clear error on interleaved multimodal placeholder overcount (#53999)

Hung Hoang (黄兴） committed about 2 months ago

[Rocm][CI] add dockerfile.xpu to rocm ci artifact (#53949)

Charlie Fu committed about 2 months ago

[Tools][Recipes] Improve sweep recommendations and short-alias parsing (#53946)

Louie Tsai committed about 2 months ago

[Kimi K3][Kernel] Support DS conv-state layout in fused KDA decode kernel (#53396)

Canlin Guo committed about 2 months ago

[Benchmark] Warn on warm prefix cache for random serve runs (#53920)

Zupeng Wang committed about 2 months ago

Bugfix: use PCP slot mappings for PIECEWISE capture (#53869)

Qiu Chunshuo committed about 2 months ago

[XPU][Dockerfile] Update UCX install (#53817)

liuzhenwei committed about 2 months ago

[Model Runner V2][Spec Decode] Skip DP sync before EAGLE/MTP draft prefill (#53694)

Giancarlo Delfin committed about 2 months ago

[warmup] overlap renderer warmup and engine core initialization (#52764)

Ning Xie committed about 2 months ago

[EC Connector] EC Offloading Connector use events instead of StepTracker (#49994)

omerpaz95 committed about 2 months ago

[Bugfix] Release CUDA graph profiling memory before KV cache allocation (#53955)

Wei Zhao committed about 2 months ago