[Mypy] Fix mypy typing for Qwen and Qianfan models (#58046)

Ashraf Bhuiyan committed 26 days ago

[Bugfix][Frontend] Detect Anthropic inline-system merge against the resolved chat template (#58754)

Robert Shaw committed 26 days ago

[Bugfix] Fix Anthropic Thinking Disabled with P/D (#58786)

Robert Shaw committed 26 days ago

[Perf][MoE] Index expert mapping lookups in RoutedExperts.load_weights (#58720)

Willian Z committed 26 days ago

[Refactor] Remove dead tests utils (#58803)

Wentao Ye committed 26 days ago

[Elastic EP] Fix EPLB load statistics during scaling (#58473)

Itay Alroy committed 26 days ago

[Security] Gate per-request multimodal processor kwargs (#58830)

Juan Pérez de Algaba committed 26 days ago

[CI] Split (B200) Miscellaneous Kernels into mHC, FLA Ops and Misc named jobs (#58609)

Thang Nguyen committed 26 days ago

[Spec Decode] Enable Gemma4 DSpark adaptive verification with FlashInfer (#57263)

qizixi committed 26 days ago

[CI] Only isolate the registry tests that need a fresh process (#58764)

Aarushi Jain committed 26 days ago

[Bugfix][Frontend][Rust Frontend] Update DeepSeek V4.1 Flash reasoning effort mappings (#58316)

zhec committed 26 days ago

[Bugfix][ROCm] AMD-Quark mixed-precision DeepSeek-V4.1 support (#57071)

xiao-llm committed 26 days ago

[RL] Add sharding-aware NCCL M2N weight transfer (#51520)

Ke Wen committed 26 days ago

[Perf][DSv4.1] Shard the Engram wkv projection across TP ranks (#58678)

Shuolei Wang committed 26 days ago

[Core] Bound draft-token RPC waits by the execute-model timeout (#58779)

Chandra Mourya committed 26 days ago

[Bugfix][Frontend] Fix Inkling tool name leaking into content after reasoning (#58792)

Baljinder Hothi committed 26 days ago

[CI] Split (H200 MIG/MI300) Basic Correctness into named jobs (#57054)

Thang Nguyen committed 26 days ago

[Bugfix][Reasoning] Count Kimi K3 reasoning tokens (#58372)

Elvir Crnčević committed 26 days ago

[Bugfix] V1: fix allowed_token_ids_mask aliasing in InputBatch.swap_states (#48419)

Rui Zhu committed 26 days ago

fix: perf: use startswith(x, i) instead of string slicing to avoid O(N^2) (#52580)

Ricardo-M-L committed 26 days ago