[kv_offload] Skip scratch groups (#57145)

Itay Etelis committed about 1 month ago

[Bugfix][KV Offload] Register the offload region in chunks (#51081)

drakosha committed about 1 month ago

[Qwen3.8-Flash-Next] Enable FP8 TP with FlashInfer TRTLLM MoE (#55867)

Canlin Guo committed about 1 month ago

[ROCm][CI] Stage H gating and MI355 test reallocation (#57080)

Andreas Karatzas committed about 1 month ago

[CPU] Align scheduler and NIXL CPU affinity per local rank (#53636)

Tianmu Li committed about 1 month ago

[CPU][s390x] Pin protobuf to 7.36.1 and drop C++ extension removal workaround (#54978)

coderfornow committed about 1 month ago

[MM][V2] Enable encoder-only ViT CUDA graph capture (#56922)

jiangkuaixue123 committed about 1 month ago

[Bugfix] Isolate supplemental FlashInfer BF16 autotuning (#57285)

Jiahan Chang (Cyrus) committed about 1 month ago

[ROCm] Stage large pageable H2D copies instead of registering them (#56343)

JohnQinAMD committed about 1 month ago

[Perf][DSpark] Stack DeepSeek V4 context WKV projections (#54674)

leo committed about 1 month ago

[Perf][Model] Qwen4Exp QSA: sm_90 tuning table for _select_config (#57273)

Shuolei Wang committed about 1 month ago

[Perf] Reuse MoE workspace for DeepGEMM warmup (#57268)

Lucas Wilkinson committed about 1 month ago

[ROCm][Perf] Enable HCA dual-stream overlap for DeepSeek-V4 (#56853)

Shanshan Shen committed about 1 month ago

Revert "[Frontend] Omit absent fields from /inference/v1/generate stream chunks" (#57302)

aoshen02 committed about 1 month ago

[Frontend] Omit absent fields from /inference/v1/generate stream chunks (#57264)

aoshen02 committed about 1 month ago

[XPU] Enable XPU eplb (#44987)

Qiming Zhang committed about 1 month ago

[Feature][Multimodal] Add cross-encoder caching to Mooncake P2P (#56242)

Zhou ziheng committed about 1 month ago

[ROCm][Bugfix] Reduce CUDA graph divergences (#57229)

Matt committed about 1 month ago

[CI] Fix ruff docstring violations in EC connector files (#57283)

Kevin H. Luu committed about 1 month ago

[A2A] Allow EPLB + SharedExpert Overlap + DeepEPv2 (#57236)

Robert Shaw committed about 1 month ago