 Karen Chung
|
986e217158
|
[Feature] Add first-class KV hints request envelope for programmatic KV management (#53423)
Signed-off-by: Karen Chung <karenc@nvidia.com>
|
2026-09-21 09:56:11 -07:00 |
|
|
Tuukka Sarviandlarryli2-amd
|
e34685dfc0
|
[ROCm][DSv4] Enable DSpark adaptive verification (#52362)
Signed-off-by: Tuukka Sarvi <tuukka.sarvi@amd.com>
Signed-off-by: larryli2-amd <larryli2@amd.com>
Co-authored-by: larryli2-amd <larryli2@amd.com>
|
2026-09-21 10:03:41 -05:00 |
|
|
Cyrus Leung
|
c8824a1e42
|
[MM] Move supports_multimodal_inputs and cache out of registry (#57913)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
|
2026-09-21 07:43:07 -07:00 |
|
|
|
0b7f11a1ee
|
[AuxOutput] Add block-keyed storage for routed-expert outputs (#45635)
Signed-off-by: aoshen02 <aoshen@inferact.ai>
Signed-off-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: aoshen02 <aoshen@inferact.ai>
Co-authored-by: OpenAI Codex <codex@openai.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Codex <noreply@openai.com>
|
2026-09-21 10:25:05 +00:00 |
|
|
Woosuk KwonandOpenAI Codex
|
97dc6b19d2
|
[Perf][Attention] Avoid redundant sparse attention metadata operations (#57885)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Co-authored-by: OpenAI Codex <noreply@openai.com>
|
2026-09-21 00:08:43 -07:00 |
|
|
|
3c84ad13ed
|
[Bugfix] Fix xgrammar feature gate bypass when JSON Schema type is a list (#48416)
Signed-off-by: john <6712432+CaiJohn@users.noreply.github.com>
Co-authored-by: john <6712432+CaiJohn@users.noreply.github.com>
Co-authored-by: Codex <noreply@openai.com>
|
2026-09-21 09:59:20 +03:00 |
|
|
liuzhenweiandKunshang Ji
|
6654bcbfca
|
[MRV2][XPU] use xpu sample kernel in mrv2 sampler (#57277)
Signed-off-by: zhenwei-intel <zhenwei.liu@intel.com>
Co-authored-by: Kunshang Ji <kunshang.ji@intel.com>
|
2026-09-21 13:24:41 +08:00 |
|
|
czhu-cohere
|
3918f3c5a3
|
[Profiler] Unify platform-aware torch profiling (#57460)
|
2026-09-20 19:17:23 -07:00 |
|
|
|
1185254851
|
[Bugfix][KV Connector] Propagate cache reset failure during sleep (#54581)
Signed-off-by: CrazyCoder <18235787078@163.com>
Signed-off-by: aoshen02 <aoshen@inferact.ai>
Co-authored-by: aoshen02 <aoshen@inferact.ai>
Co-authored-by: Codex <noreply@openai.com>
|
2026-09-21 09:25:02 +08:00 |
|
|
Yifan QiaoandClaude Fable 5.1
|
d2983f2f16
|
[Bugfix][Spec Decode] Stop dummy draft decode steps from writing KV through stale block-table rows (#56734)
Signed-off-by: Yifan Qiao <yifanqiao@inferact.ai>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
|
2026-09-21 01:06:25 +00:00 |
|
|
roikoren755andmergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
|
49ee12d742
|
[Misc] Rename --enable-mamba-fine-grained-prefix-cache (#53945 follow-up) (#57382)
Signed-off-by: Roi Koren <roik@nvidia.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
|
2026-09-20 09:50:34 -07:00 |
|
|
|
f648eed23d
|
[Bugfix][KV Offload] Skip non-prefix-cacheable groups in SimpleCPUOffload (GLM-5.3-Flash kpool tail and QSA) (#56810)
Signed-off-by: Jared Wen <jaredwen@inferact.ai>
Signed-off-by: Jared Wen <w13431838023@gmail.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Zhewen Li <zhewenli@inferact.ai>
|
2026-09-20 10:58:55 +00:00 |
|
|
|
1b9fa3eaa8
|
[Perf][GLM] Fuse the kpool tail slot mapping into one Triton kernel (#57534)
Signed-off-by: Jared Wen <jaredwen@inferact.ai>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
|
2026-09-20 18:20:55 +08:00 |
|
|
Misha GoinandErinYin
|
e5fce7b56b
|
[Core][Kernel] Share persistent workspaces for Marlin and Humming (#57421)
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: ErinYin <liang.yin@daocloud.io>
|
2026-09-20 17:28:33 +08:00 |
|
|
Mi Jiazhiandjiazhimi
|
e378275a8f
|
[Model] Pass intermediate_tensors to the model when capturing CUDA gr… (#57745)
Co-authored-by: jiazhimi <jiazhimi@tencent.com>
|
2026-09-19 23:23:29 -07:00 |
|
|
Matt MastracciandClaude Fable 5.1
|
601e1c7c51
|
[Perf] Give a prefill-only batch the model state's number of logit rows (#57416)
Signed-off-by: Matt Mastracci <matthew@mastracci.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
|
2026-09-20 04:58:30 +00:00 |
|
|
|
92b40f5a14
|
[CPU][Perf] refactor paged attention for Arm CPUs (#56045)
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: Li, Jiang <jiang1.li@intel.com>
|
2026-09-20 12:25:13 +08:00 |
|
|
|
5302d1fe40
|
[Model Runner V2] Support custom logits processors (#56497)
Signed-off-by: David Cheung <d7cheung@gmail.com>
Signed-off-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
|
2026-09-19 17:46:49 +00:00 |
|
|
Cyrus Leung
|
751f6807d9
|
[Refactor] Use dynamic MM cache in processor (#57674)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
|
2026-09-19 15:38:00 +00:00 |
|
|
Jiang WuandJiang Wu
|
6af44e2887
|
[Bugfix][CPU] Fix DeepSeek V4.1 import without Triton (#57654)
Signed-off-by: Jiang Wu <245355730+JiangLLM@users.noreply.github.com>
Co-authored-by: Jiang Wu <245355730+JiangLLM@users.noreply.github.com>
|
2026-09-19 05:46:13 +00:00 |
|
|
|
90531daf6d
|
[Core][Frontend] Bind KV-event publishers at port 0 and expose the bound endpoints (#55844)
Signed-off-by: aoshen02 <aoshen02@users.noreply.github.com>
Co-authored-by: mukeshbhandarkar <shell00035@gmail.com>
Co-authored-by: aoshen02 <aoshen02@users.noreply.github.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: aoshen02 <aoshen@inferact.ai>
|
2026-09-18 18:11:05 -07:00 |
|
|
waizuichougou
|
a1bf8ac12d
|
[Bugfix][MRV2] Match fast-prefill padding to active LoRA batches (#56456)
Signed-off-by: waizuichougou <2082431897@qq.com>
Signed-off-by: waizuichougou <105572299+waizuichougou@users.noreply.github.com>
|
2026-09-18 22:13:23 +00:00 |
|
|
|
1dc2d854c1
|
[Feat][Model] Support encoder-side SWA-bounded replay for DeepSeek-V4.1-Flash (#56227)
Signed-off-by: Yifan Qiao <yifanqiao@inferact.ai>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Woosuk Kwon <woosuk@inferact.ai>
|
2026-09-18 15:01:24 -07:00 |
|
|
|
62af3df734
|
[Bugfix][MLA] Reserve sparse prefill buffers before KV cache sizing (#57575)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
|
2026-09-18 21:22:18 +00:00 |
|
|
bnellnmandItay Etelis
|
23e26e0588
|
[KV Offloading] Back-pressure detection and remediation (#50045)
Signed-off-by: Bill Nell <bnell@redhat.com>
Signed-off-by: bnellnm <49004751+bnellnm@users.noreply.github.com>
Co-authored-by: Itay Etelis <92247226+Etelis@users.noreply.github.com>
|
2026-09-18 14:47:48 -04:00 |
|
|
frida-andersson
|
71fc70d3ae
|
[ROCm][DSv4] Fix sparse-indexer logits collapse on gfx950/gfx942 (#50455)
Signed-off-by: Frida Andersson <fanderss@amd.com>
Signed-off-by: frida-andersson <fanderss@amd.com>
|
2026-09-18 16:57:44 +00:00 |
|
|
Nicolò Lucchesi
|
a7156060c1
|
[Core] Make parallel sampling (n>1) reqs admission atomic (#53936)
Signed-off-by: NickLucche <nicolo.lucchesi@mistral.ai>
|
2026-09-18 18:08:44 +02:00 |
|
|
Yuwen ZhouandLi, Jiang <jiang1.li@intel.com>
|
b346479065
|
[CPU] Add CPU FP8 W8A8 linear/MoE support (#49942)
Signed-off-by: yuwenzho <yuwen.zhou@intel.com>
Co-authored-by: Li, Jiang <jiang1.li@intel.com>
|
2026-09-18 17:59:42 +08:00 |
|
|
olka-amdandClaude Opus 5
|
32636580a6
|
[Bugfix][MLA] TritonMLA: fix illegal memory access on causal multi-token decode (#51065)
Signed-off-by: Oleksandr Kachur <oleksandr.kachur@amd.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
|
2026-09-18 03:49:53 -05:00 |
|
|
Micah WilliamsonandAndreas Karatzas
|
d5f0a6e829
|
[Bugfix][DBO] Fix DeepEP low-latency profiling crash with DP+EP+DBO (#57502)
Signed-off-by: Micah Williamson <micah.williamson@amd.com>
Co-authored-by: Andreas Karatzas <akaratza@amd.com>
|
2026-09-18 06:33:41 +00:00 |
|
|
Juntian LiuandWoosuk Kwon
|
bf13ecc2f7
|
[Perf][MRV2] Share token-to-request mappings across KV cache groups (#57102)
Signed-off-by: Juntian Liu <Juntianl777@gmail.com>
Co-authored-by: Woosuk Kwon <woosuk@inferact.ai>
|
2026-09-18 03:14:36 +00:00 |
|
|
Simon Danielsson
|
70df48dc3d
|
[Bugfix][KV Cache][GLM-5.3-Flash] Disable slot mapping kernel for the kpool tail buffer (#57317)
Signed-off-by: simondanielsson <simon.danielsson99@hotmail.com>
|
2026-09-17 22:42:54 -04:00 |
|
|
Ren Yuzhou
|
7b94293627
|
[Bugfix][ROCm][KV Offload] Use private pinned tensors for CPU KV offload (#57160)
Signed-off-by: Yuzhou Ren <ryzcqu@outlook.com>
|
2026-09-18 02:35:23 +00:00 |
|
|
|
092bdd6d57
|
[Frontend][Core] Add release_kv_cache_memory() API (#44890)
Signed-off-by: Dakai An <dakaian108@gmail.com>
Signed-off-by: AlanFokCo <alanfok2868@gmail.com>
Signed-off-by: aoshen02 <aoshen@inferact.ai>
Signed-off-by: Dakai An <77474977+andakai@users.noreply.github.com>
Co-authored-by: AlanFokCo <alanfok2868@gmail.com>
Co-authored-by: aoshen02 <aoshen@inferact.ai>
Co-authored-by: OpenAI Codex <noreply@openai.com>
|
2026-09-18 07:12:12 +08:00 |
|
|
Wei Zhao
|
ccde9c61e9
|
[Bugfix] Restore KV cache metadata GET method for external event consumer (#56925)
Signed-off-by: Wei Zhao <51183510+wzhao18@users.noreply.github.com>
|
2026-09-17 23:01:35 +00:00 |
|
|
|
e0050f287a
|
[Bugfix][ROCm] Add record_logical_topk_ready to ROCMAiterMLASparseImpl (GLM-5.3-Flash boot crash) (#57252)
Signed-off-by: Mustafa YILDIRIM <mustafa@character.ai>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Andreas Karatzas <akaratza@amd.com>
|
2026-09-17 22:34:56 +00:00 |
|
|
Woosuk KwonandCodex
|
80447d2765
|
[Bugfix][DSv4.1] Fix FlashInfer DSpark non-causal attention (#57432)
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Co-authored-by: Codex <noreply@openai.com>
|
2026-09-17 15:14:37 -07:00 |
|
|
Maroon AyoubandRobert Shaw
|
9612f77077
|
[Bugfix] Release stale FlashMLA workspace views after growth (#56902)
Signed-off-by: Maroon Ayoub <mayoub@redhat.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
|
2026-09-17 16:23:15 -04:00 |
|
|
|
ac2f0ea82c
|
[MoE][Bugfix] Skip SP padded rows in grouped MoE routing (#56079)
Signed-off-by: Elvir Crncevic <elvircrn@gmail.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
|
2026-09-17 19:36:19 +00:00 |
|
|
|
67e5b0acc9
|
[Bugfix][HiSparse][NIXL] Import full blocks without tail prefill on D (#57049)
Signed-off-by: NickLucche <nicolo.lucchesi@mistral.ai>
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: OpenAI Codex <noreply@openai.com>
Co-authored-by: Lucas Wilkinson <lwilkins@redhat.com>
|
2026-09-17 15:26:10 -04:00 |
|
|
Michele Campi
|
6a344d8c85
|
[Bugfix][Metrics] Do not log a 0.0% prefix cache hit rate before any query (#54990)
Signed-off-by: Michele Campi <215741962+MicheleCampi@users.noreply.github.com>
|
2026-09-17 18:19:13 +00:00 |
|
|
Sriram KumarandDouglas Lehr
|
ff6b5808c4
|
[ROCm] Resolve the indexer fp8 cache dtype once at import (#53792)
Signed-off-by: Sriram Kumar <sriramkumar.kishorekumar@amd.com>
Co-authored-by: Douglas Lehr <91553416+dllehr-amd@users.noreply.github.com>
|
2026-09-17 17:51:35 +00:00 |
|
|
HieDeanandhiedean
|
493d4b3c04
|
[bugfix] Mark draft tokens to rebuilt their embeddings. (#57356)
Signed-off-by: HieDean <799287043@qq.com>
Co-authored-by: hiedean <hiedean@tju.edu.cn>
|
2026-09-17 15:20:37 +00:00 |
|
|
|
9e1cd262cb
|
[Bugfix] Use DP index for dense DP weight updates and EC CPU region (#56950)
Signed-off-by: lxy-alexander <lxy_alexander@outlook.com>
Signed-off-by: aoshen02 <aoshen02@users.noreply.github.com>
Co-authored-by: aoshen02 <aoshen02@users.noreply.github.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: aoshen02 <aoshen@inferact.ai>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
|
2026-09-17 22:44:36 +08:00 |
|
|
Robert ShawandClaude Opus 5
|
861a299adc
|
[Bugfix][Model Runner V2] Route dummy tokens to MoE experts during profiling (#57270)
Signed-off-by: Robert Shaw <robertgshaw2@gmail.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
2026-09-17 10:11:14 -04:00 |
|
|
CaiJohnandjohn
|
75c71390d5
|
[Bugfix] Escape control characters in xgrammar choice grammar (#48115)
Signed-off-by: john <6712432+CaiJohn@users.noreply.github.com>
Co-authored-by: john <6712432+CaiJohn@users.noreply.github.com>
|
2026-09-17 15:57:47 +03:00 |
|
|
Martín el CheikhandClaude
|
2e50824766
|
[Bugfix] Fall back to native sampling when FlashInfer cannot target the GPU (#48956)
Signed-off-by: martin el cheikh <martinelcheikh@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
|
2026-09-17 12:27:54 +00:00 |
|
|
aoshen02andaoshen02
|
9ff08f9493
|
[Bugfix] Honor skip_reading_prefix_cache for KV connector hits (#57269)
Signed-off-by: aoshen02 <aoshen02@users.noreply.github.com>
Co-authored-by: aoshen02 <aoshen02@users.noreply.github.com>
|
2026-09-17 03:46:28 -07:00 |
|
|
|
40b40d1d39
|
[Bugfix][KV Offload] Register the offload region in chunks (#51081)
Signed-off-by: Mikhail Kostryukov <mike@triptrack.net>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Colton Ottley <coltonottley@gmail.com>
Co-authored-by: Or Ozeri <oro@il.ibm.com>
|
2026-09-17 13:19:56 +03:00 |
|
|
jiangkuaixue123andCodex
|
de24e51908
|
[MM][V2] Enable encoder-only ViT CUDA graph capture (#56922)
Signed-off-by: jiangkuaixue123 <jiangxiaozhou111@163.com>
Co-authored-by: Codex <noreply@openai.com>
|
2026-09-17 16:12:17 +08:00 |
|