Commit Graph

Select branches

Hide Pull Requests

main

patches

sm86

Mono Color

• d3130d878c [CI] Pin GitHub Actions to commit hashes in macos-smoke-test.yml (#38290) Russell Bryant 2026-06-25 09:48:44 -04:00
• 9bfd878a48 [MoE] [MoE Refactor] Add moe kernel oracle abc 37753 (#43461) Qiuyang Yueandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-25 06:34:03 -07:00
• 2365b7a8e7 [Hardware][AMD][CI] Mirror Basic Models (Others) and Weight Loading Multiple GPU test groups (#46668) Matt 2026-06-25 08:25:09 -05:00
• 15be78732b [NIXL][Mamba] Add Mamba1 support to NIXL P/D disaggregation (#45019) Asaf Gardinandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-25 15:50:41 +03:00
• 92221485aa [CPU][CI/Build] Allow more CPU CI agents (#46702) Li, Jiangandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-25 19:33:39 +08:00
• a6f41ab678 [XPU][CI]Refine .buildkite/ci_config_intel.yaml for Intel GPU CI (#46674) xiangdong 2026-06-25 16:58:26 +08:00
• c63cd4906c [ROCm][ [Perf] sparse attention optimization on minimax-m3 (#46546) 4 people 2026-06-25 04:56:00 -04:00
• 638b1a99cc [CPU][RISC-V] Add RVV path for W4A8 INT4 GEMM (#45269) 3 people 2026-06-25 16:18:10 +08:00
• 72adb20a6a [Model] Remove AquilaForCausalLM, AquilaModel (#46605) Tiezhen WANGandClaude 2026-06-25 16:08:26 +08:00
• 2396d91e93 [CPU][Spec Decode] Enable DFlash SD for CPU (#44029) guybdandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-25 10:32:48 +03:00
• 9b215ae60b [Rust Frontend] Forward VLLM_ENGINE_READY_TIMEOUT_S via --args-json (#44610) Kaiand图灵 2026-06-25 15:25:08 +08:00
• 4d3b4b9b01 [Rust Frontend] Make ToolParserOutput a seq of ToolParserEvent to preserve order (#46584) Bugen Zhao 2026-06-25 14:27:07 +08:00
• 77c1d9fe9b [ROCm][Perf] Tune wvSplitK on gfx1151 (#40784) Matthias Gehre 2026-06-25 08:17:46 +02:00
• 36fd7e8b86 [SimpleCPUOffloadConnector] Fix remaining global→block conversions under PCP/DCP (#46394) Jeff (Junze) Ma 2026-06-24 23:05:24 -07:00
• fc61c6fc26 [Perf] Enable + tune FlashInfer fused allreduce at world_size=16 on SM 10.3 (GB300) (#46392) Jeff (Junze) Maandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-24 23:04:17 -07:00
• e2af449c39 [Hardware][AMD][CI] Move Metrics, Tracing (2 GPUs) & make optional (#46686) Matt 2026-06-25 00:49:33 -05:00
• 3f5a1e1733 [ROCm][CI] Expand basic correctness target suites (#46573) 3 people 2026-06-24 23:18:57 -05:00
• 710ebaa189 [ROCm][Bugfix] Fix chunk alignment when using context parallelism with TRITON_MLA (#46114) 3 people 2026-06-24 23:07:28 -05:00
• 1aad125815 [CPU] Enable chunked prefill and prefix caching for qwen3.5 (#46202) 3 people 2026-06-24 20:49:21 -07:00
• dc55936f64 [AMD][CI] Fix Pipeline + Context Parallelism test group (#46650) Aarushi JainandAndreas Karatzas 2026-06-24 22:23:42 -05:00
• 76c3c4ff63 [Rust Frontend] Introduce unified parser interface & combined parser (#46583) Bugen Zhao 2026-06-25 11:17:31 +08:00
• efb5acffd5 [Bugfix] fix: stream Mimimax m2 tool call string arguments (#46382) 3 people 2026-06-25 11:12:45 +08:00
• 6e3a983cf3 [ROCm] Remove erroneous inclusion of gptq_marlin as supported quant scheme on ROCm (#46655) Micah WilliamsonandAndreas Karatzas 2026-06-24 21:27:19 -05:00
• 1273a8f05a [Kernel] Add swap AB optimization to fused_moe_kernel (#36559) Xin Yang 2026-06-24 18:44:30 -07:00
• 9e88e969c0 [Perf][KVConnector][Mooncake] Parallelize KV load with a receive-thread pool (#45971) 3 people 2026-06-24 18:25:12 -07:00
• dda3aca47f [Speculative Decoding] Propagate norm_output and fc_norm config for Eagle3 speculators (#46488) 3 people 2026-06-25 02:51:33 +02:00
• 23aed9b0ee [Kernel] Enable PDL for per_token_group_quant_8bit_kernel (#46508) Jee Jee Li 2026-06-25 08:42:51 +08:00
• cd347298e8 [Frontend] Port seed_oss to the streaming parser engine as a Qwen3 subclass (#46314) Maxwill Lin 2026-06-24 17:08:42 -07:00
• b69816043a [Bugfix][MooncakeStore] track resumed requests via scheduler's resumed_req_ids (#46595) Yifan Qiao 2026-06-24 16:50:56 -07:00
• fc7fc421e9 [Kernel][MoE] Allow FlashInfer MXINT4 MoE for gated SiLU (#46518) Kaihang Jiang 2026-06-24 19:32:50 -04:00
• e06a83445c [Bugfix] Normalize slashes in Helion GPU names (#46101) cyq 2026-06-25 07:22:49 +08:00
• d7ab9be775 [Bugfix] Support -1 (invalid/non-local) slots in topk_ids for Triton MoE (#46408) Woosuk KwonandClaude 2026-06-24 13:59:42 -07:00
• 6a1570711c [Bugfix] Support non-power-of-2 top_k in legacy triton_kernels routing (#46406) Woosuk KwonandClaude 2026-06-24 13:52:09 -07:00
• d6696e2385 [ROCm] Begin Deprecation Window for CUDA_VISIBLE_DEVICES on ROCm (#46636) Micah Williamson 2026-06-24 15:40:28 -05:00
• 84c2f9f0fb [Frontend] Fix Kimi K2 tool call IDs for required tool choice (#46344) Chauncey 2026-06-25 03:59:40 +08:00
• 49f2104c53 [Feature] Support DCP with FP8 KV cache in MLA decode path (#44044) 3 people 2026-06-24 12:28:17 -07:00
• d511b5bae9 Chore: Fix minor doc sentence, grammar, quote errors (#40469) Ashwin PhadkeandHarry Mellor 2026-06-25 00:28:24 +05:30
• 3c43237233 [Bugfix][Model Runner V2][Spec Decode] Fix int32 offset overflow in sampler kernels (#46560) 3 people 2026-06-25 02:00:57 +08:00
• 56ca5997ea Humming support for 2/3/5/6/7-bit pack-quantized weight-only inference (#46389) HDCharlesandClaude Opus 4.6 2026-06-24 13:53:54 -04:00
• cf57311187 Run DeepSeek-V2-Lite prefetch-offload eval eager on ROCm (#46386) Aarushi Jain 2026-06-24 12:25:57 -05:00
• e7df232288 [KV Offload] Gate packed HMA KV cache on cross-layer config (#46252) Lucas Wilkinson 2026-06-24 11:55:30 -04:00
• b3a688cb9e [ROCm] Fix OOB During Model Warmup With ROCM_ATTN and MRV2 (#46548) Micah WilliamsonandMatthew Wong 2026-06-24 10:53:21 -05:00
• 1cd3e0e945 [Bug] Fix IndentationError: expected an indented block after 'with' statement (#46627) Wentao Ye 2026-06-24 11:14:17 -04:00
• f889325c51 [KV Offload] Use background thread for mmap / cpu_tensors pinning (#45850) Yiwei Hu 2026-06-24 23:13:27 +08:00
• bb61177e49 [KV Offloading] Replace bool|None lookup return with LookupResult enum (#46363) Ronen SchafferandOr Ozeri 2026-06-24 18:06:08 +03:00
• 7f99e80c3b [Perf][ThinkingBudget] reduce search space for thinking tokens (#46425) Walter Beller-Moralesandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-24 11:02:25 -04:00
• 2801b11156 [Test] Pin block_size in auto-fit max_model_len test (#45914) Liangliang Maandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-24 22:56:21 +08:00
• 007b5a52ed [Log] Update to log once (#46511) Wentao YeandTJian 2026-06-24 10:45:16 -04:00
• 24d5186138 [Bugfix] Re-enable FP8 MoE on NVIDIA Thor (#46339) Cyrus Leung 2026-06-24 22:35:46 +08:00
• 7dc036058b [Doc] Document Qwen3.6 (dense + MoE) ViT CUDA graph support (#44720) Nemani Harsha Vardhan 2026-06-24 16:35:08 +02:00
• 61ee183d28 [ROCm] Fix AITER FP8 quantization schema tests (#46414) djramicandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-24 16:29:19 +02:00
• 84c62e1cbd [Model Runner V2][MM] Support EVS (#46535) Nick Hillandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-24 07:18:56 -07:00
• 061043eaca [CPU][Perf] Accelerate unquantized MoE for AArch64 (#46353) Fadi Arafeh 2026-06-24 15:14:35 +01:00
• 93ec645878 [Bugfix] Fix illegal memory access from a forward during a partial wake_up (#44483) 3 people 2026-06-24 22:12:23 +08:00
• 563c628968 [XPU] bump up vllm_xpu_kernels to v0.1.10.1 (#46607) Kunshang Ji 2026-06-24 22:05:31 +08:00
• 0bc479e6eb [Perf][LoRA] Replace O(n) list.index() with a dict in convert_mapping (#46542) LynnandClaude Opus 4.8 2026-06-24 08:41:46 -05:00
• 62890e204c Fix duplicated logging when loading a corrupt or partial video (#46467) Tae Jeong 2026-06-24 22:14:13 +09:00
• a2cb08b3d5 [Misc][PD] Disable bidirectional xfer mode for NixlPushConnector (#46473) Nicolò Lucchesi 2026-06-24 15:14:05 +02:00
• cf9fd6457e Fix KV offload request-finished lifecycle contract (#46284) Rui YinandOr Ozeri 2026-06-24 20:42:40 +08:00
• d4448b511d [XPU][Docker] switch to ubuntu 24.04 as base image (#45973) Kunshang Ji 2026-06-24 20:39:20 +08:00
• f1a6703edd [Bugfix][Config] Keep pydantic validation for fields with a TYPE_CHECKING Literal alias (#46220) Ting SUNandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-24 19:25:50 +07:00
• 160c80a34c [Rust Frontend] Raise frontend JSON body limit (#46582) Roy Wang 2026-06-24 20:15:31 +08:00
• f237e16b41 [KV Offload] Replace OffloadingHandler with OffloadingWorker (#45053) Martin Hickey 2026-06-24 12:44:24 +01:00
• 70749fdcca [Feature] Triton INT4 per-token-head KV cache quantization (#40835) JartXandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-24 12:21:25 +02:00
• d20dbf921b [Mooncake] Only check and store new KV cache range (#46412) Wei ZhaoandYifan Qiao 2026-06-24 06:10:50 -04:00
• ede54b926e set AttentionCGSupport.UNIFORM_BATCH for fa2 on xpu (#46555) Xinyu ChenandKunshang Ji 2026-06-24 18:05:02 +08:00
• 52fbe12283 [Perf][Multimodal] Avoid building a full timestamps list in video frame sampling (#46543) LynnandClaude Opus 4.8 2026-06-24 04:38:27 -05:00
• dc0d318177 [Attention] Add FLASH_ATTN_MLA_SPARSE backend for Hopper sparse MLA (#46189) Dakai An 2026-06-24 17:33:10 +08:00
• d7c1821b5a [Model][MiniMax-M3] Add pipeline parallelism support (#45810) soaringk 2026-06-24 16:23:03 +08:00
• 4cd1a84c88 [Model] Remove BaiChuanForCausalLM and BaichuanForCausalLM (#46362) Tiezhen WANGandClaude 2026-06-24 16:13:57 +08:00
• 191826ec61 [CI/Build] Fix topk histogram build on SM75 (#46550) Mohammad Miadh Angkad 2026-06-24 15:51:11 +08:00
• 549c7074cd [ROCm][CI] Skip the MoE Marlin tile-padding helper assertion (#46580) Andreas Karatzas 2026-06-24 02:31:33 -05:00
• 489abadfb8 feat: support to OpenMOSS-Team (#44124) 4 people 2026-06-24 15:08:13 +08:00
• 96de8bb389 [MoE] Free unused MXFP4 scales in OAI Triton Backend (#46549) Woosuk Kwon 2026-06-24 00:06:41 -07:00
• 9d6fdc2901 [Kernel] GLM5 Router GEMM (#46385) Jee Jee Li 2026-06-24 13:54:50 +08:00
• 4c5bc41ba6 [Bugfix][Spec Decode] Fix probabilistic sampling for parallel drafting (#45956) Benjamin Chislett 2026-06-24 01:36:23 -04:00
• ac1fa74616 [Bugfix] Fix NemotronLayerNorm1P hardcoded cuda device type (#46495) Michał Ganczarenko 2026-06-24 08:21:02 +03:00
• 556bc4e3a0 Upgrade tpu-inference to v0.23.0 (#46568) Sting Lin 2026-06-24 12:15:14 +08:00
• 05a0caba91 [Mooncake] Optimize lookup pool key string construction (#46188) Wei Zhao 2026-06-23 23:51:53 -04:00
• 7ee4d22009 [Spec Decode] Reject placeholder (-1) draft tokens in rejection sampler (#46533) Nick Hill 2026-06-23 20:32:32 -07:00
• ce9f64020b [Rust Frontend] Pass effective reasoning_parser_kwargs for structured output (#46360) Bugen Zhaoandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-24 11:13:44 +08:00
• 4ed8eaafb0 [Rust Frontend] Integrate xgrammar-structural-tag for strict and required tool calling (#46057) Bugen Zhaoandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-24 10:46:49 +08:00
• 6af0559ddb [Core][DP] Throttle prefills based on local prefill work (#46532) Nick Hillandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-23 19:27:12 -07:00
• e2bdc24612 [ROCm][Bugfix] Fix use_v2_model_runner inside Ray driver thread (#45998) Micah Williamsonandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-23 19:41:56 -05:00
• bcbeaac786 [ROCm][CI] Stage C-II of gating additional test groups (#46537) Andreas Karatzas 2026-06-23 19:36:40 -05:00
• e48f2aa4ca [Bugfix][Frontend] Emit a content block for empty Anthropic completions (#46525) Maxwill Lin 2026-06-23 17:04:26 -07:00
• d86c66c981 [Feat] Add runtime monitor for post-warmup CuTeDSL compilation (#46167) Roberto L. Castro 2026-06-24 01:33:17 +02:00
• 80e511772f [ROCm][Bugfix][Perf] enable shared expert fusion for Qwen3.5 (#44434) Nico Holmberg 2026-06-23 16:19:51 -07:00
• 855cd4d787 [Perf][DSv4/DSv3.2] Add cluster-cooperative topK kernel for low-latency scenarios (#43008) Roberto L. Castro 2026-06-24 01:11:00 +02:00
• 3cc871aaf1 [Perf] Skip detokenization in online beam search (#46422) Guy StoneandClaude 2026-06-23 18:46:09 -04:00
• 0a3e2dbc09 [Optimization] Skip DP padding tokens in MoE (#46428) Woosuk Kwonandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-23 14:54:46 -07:00
• 84f13374b3 [CI] Fix test_auto_gptq on ROCm CI (#46164) fxmarty-amdandmergify[bot] <37929162+mergify[bot]@users.noreply.github.com> 2026-06-23 23:38:06 +02:00
• b28103e1ca [ROCm][CI] Shard LM Eval Qwen3-5 Models (B200-MI355) in AMD CI (#46520) Micah Williamson 2026-06-23 16:32:05 -05:00
• abc33134fa [CI Test] Mark batch invariance test flaky (#46530) Wentao Ye 2026-06-23 17:01:34 -04:00
• 6617db1bfb [Bugfix][Frontend] Emit non-ASCII tool-call arguments without \uXXXX escapes (#46308) Maxwill LinandEazyReal 2026-06-23 13:43:11 -07:00
• 899d72a58c [Bugfix][ToolParser] Handle braces in required tool streaming strings (#45389) Ting SUNandFlora Feng 2026-06-24 03:29:34 +07:00
• 11b56b2ff2 [Kernel] Add FlashInferCutedslMxfp8LinearKernel (cute-dsl mm_mxfp8) (#46393) Yongye Zhu 2026-06-23 12:45:49 -07:00
• 0d4d164488 [Bugfix] Allow flashinfer_cutlass as a clamped NVFP4 MoE backend (#46492) 3 people 2026-06-24 03:43:36 +08:00
• 0775b882ba [NVFP4 MoE/Deepseek V4] Marlin: wire SwiGLU clamp + allow it for clamped models on non-Blackwell (#45836) Mike G 2026-06-23 12:21:19 -07:00
• 7c2e08451a [Docker] Remove redundant flashinfer download-cubin step (#46517) Michael GoinandClaude 2026-06-23 13:16:51 -06:00