[Refactor] Derive is_reasoning_end from the engine grammar (#56200)

Flora Feng committed about 1 month ago

[Bugfix][Rust Frontend] Honor `add_generation_prompt` in DeepSeek V3.2/V4/V4.1 renderers (#56593)

Yeonwoo Sung committed about 1 month ago

[Perf][DSpark] Collapse DeepSeek-V4.1 draft states before SP all-gather (#56903)

Juntian Liu committed about 1 month ago

[CI] Shard V1 KV Connectors 1→4 (#56324)

Thang Nguyen committed about 1 month ago

[CI] Fix NIXL push worker test stub after failure deferral (#56955)

Kevin H. Luu committed about 1 month ago

[Frontend] Use XGrammar schema constraints for DeepSeek V4.1 (#56408)

Yixin Dong committed about 1 month ago

[Feature][PCP] Support decode-only FULL CUDA graphs (#53867)

Qiu Chunshuo committed about 1 month ago

[Bugfix][Frontend] Handle aborted requests in beam search (#56249)

shaohuaxi committed about 1 month ago

[CI] Extend Arm CPU kernel shard timeout (#56601)

Kevin H. Luu committed about 1 month ago

[Build] Move DeepGEMM pin to the vLLM fork and update the Mega MoE call convention (#56876)

Yongye Zhu committed about 1 month ago

[KV Offload] Add KVCR secondary-tier adapter (#53624)

Moein Khazraee committed about 1 month ago

[Bugfix][SparseMLA] Fix piecewise cudagraph capture crash in index group (#56825)

Jared Wen committed about 1 month ago

[Doc] Clarify ITL vs TPOT Prometheus metrics (#55283)

Jyotirmoy Roy committed about 1 month ago

[Bugfix][Kimi-K3] Keep transient checkpoints out of prefix-cache eviction (#56794)

Sy03 committed about 1 month ago

[CI] Select the supported DCP backend for PCP eval (#56677)

Kevin H. Luu committed about 1 month ago

[feature] [xgrammar] support `patternProperties`/`propertyNames`/`unevaluatedProperties` kw for object types (#42904)

cjackal committed about 1 month ago

[CI] Extend DSv4 engine readiness timeout on main (#56735)

Kevin H. Luu committed about 1 month ago

[Bugfix] Fix stale HPC QK-norm weights after weight refit (#56706)

aoshen02 committed about 1 month ago

[Cleanup] Remove vestigial `tpu_input_batch.py` (#56898)

Nick Hill committed about 1 month ago

[Model Runner V2] Tolerate lack of pinned memory support (#56908)

Nick Hill committed about 1 month ago