Commit Graph

Select branches

Hide Pull Requests

bot/update-est-time-20260928

dsv41-pd

dsv41-pd-visioncp

main

Mono Color

• ae62279071 Fix data parallel controller launch for num nodes > 2 (#12822) Lianmin Zheng 2025-11-07 12:13:21 -08:00
• 5c9273c032 [router] Support structured model output for openai and grpc router (#12431) Keyang Ru 2025-11-07 10:16:45 -08:00
• e316bcacb1 update multimodal_gen readme (#12825) MickandXinyuan Tong 2025-11-08 01:04:44 +08:00
• 0fe9c1f70b Fix piecewise cuda graph ci test (#12836) Ke Bao 2025-11-08 00:25:00 +08:00
• 61bfd9fa9b optm(checkpoint-engine): disable multi-thread loading when update weights (#12374) Yang Kaiyong 2025-11-07 21:30:04 +08:00
• c67fce160e Fix missing dp_max_padding argument in set_dp_buffer_len (#12812) Chen1022 2025-11-07 21:27:43 +08:00
• bef37d6de8 [Deepseek V3.2] Only skip Indexer logits computation when is_extend_without_speculative (#12816) hlu1 2025-11-07 01:34:04 -08:00
• bc25ea6762 [MoE] Add Comprehensive MoE Integration Tests (#12090) Jonah Bernard 2025-11-07 03:34:46 -05:00
• 1fa788ec14 [Bugfix] Fix illegal memory access (#12758) elvischenv 2025-11-07 13:59:57 +08:00
• 125f76ea44 [Test] Add DeepSeekV3.2 NSA Indexer Test Suite (#12520) Johnsonms 2025-11-06 21:28:43 -08:00
• d8736c756a fix multimodal gen issues (#12765) Yuhao YangandMick 2025-11-07 13:19:49 +08:00
• 3b1cc466c0 fixes hardcoded "cuda" device references in unit tests to use a dynamic device selection (#12761) Kalyan Kumar 2025-11-07 09:08:35 +05:30
• 0ee5ab5a74 Add PP support for dots_vlm (#12763) Tianyu Guo 2025-11-07 11:14:27 +08:00
• ed5e905c89 [HotFix]: Add missing SGLANG_EPLB_HEATMAP_COLLECTION_INTERVAL env var (#12776) hzh0425 2025-11-07 10:58:44 +08:00
• d9c812d881 [hotfix]: Resolve ModuleNotFoundError in PD deployment for is_in_ci() (#12772) hzh0425 2025-11-07 10:48:03 +08:00
• 7257525cce [DeepSeek-V3.2][NSA] Enable MHA Pathway for Short Sequence Prefill on B200 (SM100) (#12788) YAMY 2025-11-06 17:26:08 -08:00
• 34ba10efd6 [AMD] Update wave-lang to 3.8.2 (#12576) Gaurav Verma 2025-11-06 15:51:44 -08:00
• a119363f08 ignore the deepgemm check when the model weight with nvfp4 and moe ba… (#12782) Rain Jiang 2025-11-06 15:27:28 -08:00
• fb314d7bb1 Add bot-bump-kernel-version-to-sglang workflow (#12794) alisonshao 2025-11-06 14:27:40 -08:00
• 3a64844a1c Fix hang in deepgemm compilation with symmetric memory enabled (#12715) Nicolas Castet 2025-11-06 16:06:22 -06:00
• 585c417fc0 chore: update CODEOWNERS (#12795) Yineng Zhang 2025-11-06 13:11:28 -08:00
• b0d1c21d03 [fix] Only enable flashinfer all reduce fusion by default for single-node servers (#12724) Lee Nau 2025-11-06 11:53:58 -08:00
• b07c5e4080 Pin uvloop to 0.21.0 (#12279) R0CKSTAR 2025-11-07 03:33:31 +08:00
• 1772671be5 [docs] Remove redundant --disable-radix-cache option from (#12717) Rahul Chalamala 2025-11-06 11:25:49 -08:00
• 6e6009fb6b [CPU] Fix TP padding case with weight block size (#8243) jianan-gu 2025-11-07 03:24:48 +08:00
• 88a2a34041 [Fix]HTTP Stream raise exception (#11904) Jimmy 2025-11-07 03:23:12 +08:00
• 78c58621d2 [mem pool] bugfix: wrong position for self.device in Mamba (#12684) Teng Ma 2025-11-07 03:05:42 +08:00
• c3bb348dad [Docs] fix dead links in multiple documentation pages (#12764) Mattheliu 2025-11-07 02:49:32 +08:00
• 4e234b4cf9 [Intel XPU]Update pytorch xpu to 2.9 (#12363) gaopengff 2025-11-07 02:34:23 +08:00
• 4cc725ac1c [Intel]Add 'intel_xpu' attention backend for llama4 (#11051) gaopengff 2025-11-07 02:33:41 +08:00
• 2cb42dc184 [router][grpc] Add more mcp test cases to responses api (#12749) Chang Su 2025-11-06 10:20:00 -08:00
• 0c006b8809 chore: bump SGLang version to 0.5.5 (#12739) sglang-bot 2025-11-07 00:46:19 +08:00
• cd135bfe30 Update dsv3 quantization auto setting for sm100 (#12778) Ke Bao 2025-11-06 22:39:13 +08:00
• fc84b0730c [Refactor] Refactor fused_moe_triton tuning tools: extract shared utils, add EP/MLLM support, reduce overhead (#12440) 3 people 2025-11-06 20:54:42 +08:00
• 8be0e1bc9c fix ci (#12760) Keyang Ru 2025-11-06 00:21:11 -08:00
• 8e1d6756d5 docs: document video-capable multimodal models (#12565) Amit Prakash 2025-11-06 13:36:19 +05:30
• 0da30dbc20 fix qwen3-omni audio length < 30s (#12674) jiapingWandcanghua 2025-11-06 16:05:28 +08:00
• 58095cb00a Add timing metrics for requests (#12646) yinghuiandScott Lee 2025-11-05 23:07:16 -08:00
• fd3034da75 [VLM] Optimize qwen_vl preprocess_video (#12240) 3 people 2025-11-06 14:55:01 +08:00
• fbbe16faab [GDN] Fuse b.sigmoid(), fused_gdn_gating and unsqueeze into one kernel: up to 0.85% e2e speedup (#12508) Binyao Jiang 2025-11-05 22:52:04 -08:00
• 6a1a64fa41 [BUGFIX] fix output_ids in abort (#12737) Yi Zhang 2025-11-06 14:51:47 +08:00
• e2715cf871 fix mamba prefix cache leak caused by abort (#12693) Yi Zhang 2025-11-06 14:51:03 +08:00
• 74630ba310 [router][ci] Disable cache (#12752) Keyang Ru 2025-11-05 22:49:03 -08:00
• 73e9a2ef5c fix: tiny fix cli (#12744) Mick 2025-11-06 14:45:50 +08:00
• 837b08eb2e [router][grpc] Support mixin tool calls in Responses API (#12736) Chang Su 2025-11-05 21:29:45 -08:00
• bb6a21cd99 [Fix]Tiny fix in Dockerfile (#12748) Baizhou Zhang 2025-11-05 21:04:33 -08:00
• 32ec68faf9 keep attention backend document up to date (#12741) b8zhong 2025-11-05 20:41:20 -08:00
• 3c0a6df82d [chore] Fix triton installation for cu13 image (#12742) Baizhou Zhang 2025-11-05 20:20:18 -08:00
• 2104d20eba Temporarily fix missing routed_scaling_factor for CompressedTensorsWNA16MoEMethod (#12738) Atream 2025-11-06 12:01:03 +08:00
• f235498eca DeepSeek-V3.2: Add Adaptive MHA Attention Pathway for Short-Sequence Prefill (#11892) YAMY 2025-11-05 19:33:26 -08:00
• 149dc9aab1 Add nightly test multi gpu configs (#12721) alisonshao 2025-11-05 19:30:19 -08:00
• 9a954982de [chore] SGLang tag management in Dockerfile (#12734) Baizhou Zhang 2025-11-05 19:03:26 -08:00
• 97be66c358 fix sgl-kernel version (#12723) gongwei-130 2025-11-05 19:01:03 -08:00
• 74243dffa2 Revert "[router] web_search_preview tool basic implementation" (#12716) Keyang Ru 2025-11-05 18:57:39 -08:00
• 4ea4c48bb0 Revert "[ci] fix permission" (#12732) Keyang Ru 2025-11-05 18:26:29 -08:00
• cf5d27e30a [CPU] Upgrade default PT version to 2.9 (#12611) Zaili Wang 2025-11-06 10:16:35 +08:00
• 9ec6031d70 [chore]Remove dockerfile from target file of bump kernel version (#12728) Baizhou Zhang 2025-11-05 18:01:19 -08:00
• a5affb0caf [ci] fix permission (#12729) Keyang Ru 2025-11-05 17:43:48 -08:00
• 5925d3d719 fix labeler (#12718) Keyang Ru 2025-11-05 17:34:56 -08:00
• 7ef1964a66 [router] add basic ci tests for gpt-oss model support (#12651) Keyang RuandChang Su 2025-11-05 15:45:11 -08:00
• b0476a0642 [router][quick fix] Add minimal option for reasoning effort in spec (#12711) Keyang Ru 2025-11-05 15:43:19 -08:00
• ffba61a100 [router][grpc] Make harmony parser checks recipient first before channel (#12713) Chang Su 2025-11-05 15:33:52 -08:00
• 3c219eb0e4 [misc] Change sync-labels to false (#12714) Chang Su 2025-11-05 15:33:33 -08:00
• 1ffdcdc430 Revert "Commented out b200 tests due to runner shortage (#12609)" (#12712) Kangyan-Zhou 2025-11-05 15:09:49 -08:00
• c7d57d5bb3 Fix CI and style (#12658) Lianmin Zheng 2025-11-05 15:08:15 -08:00
• 80802c4cc6 [router][ci] speed up python binding to 1.5 min (#12673) Keyang Ru 2025-11-05 14:52:33 -08:00
• 83b104ee8c [misc] Add labeler for automatic labeling (#12710) Chang Su 2025-11-05 14:44:43 -08:00
• 141278048e [NVIDIA] Fix unit test of MoE and add it to nightly ci (#12709) Kaixi Hou 2025-11-05 14:33:18 -08:00
• 82f39dc11d Add mm_fp4 trtllm backend (#12406) Shu Wang 2025-11-05 14:31:46 -08:00
• 627bac649c Support Expert Deferral Mechanism in KTransformers (#12586) 3 people 2025-11-06 05:41:52 +08:00
• 3651cfbf62 [router] fix: validate HTTP status codes in health check (#12631) wyx 2025-11-06 05:37:17 +08:00
• c8547ecddd Enable Aiter Attention for VL model (#12699) Morpheus Guoandyuechguo 2025-11-06 05:01:23 +08:00
• 7bc1dae095 WIP: initial multimodal-gen support (#12484) +919 people 2025-11-06 04:28:52 +08:00
• 4fe53e5888 [router][grpc] Support streaming parsing with Tool Choice in chat completions API (#12677) Chang Su 2025-11-05 12:12:21 -08:00
• fb2e816e83 Fix server args for gpt oss so users can override the moe runner backend (#12696) Lianmin Zheng 2025-11-05 11:36:59 -08:00
• 7c45b8b4bb [CI] Fix qwen3-vl lora nightly ci (#12708) Baizhou Zhang 2025-11-05 11:00:13 -08:00
• ba5b68236f Commented out b200 tests due to runner shortage (#12609) Kangyan-Zhou 2025-11-05 10:47:43 -08:00
• 508d2f7aa2 add Kimi k2 reasoning parser (#12702) bigmoyan 2025-11-06 00:37:54 +08:00
• a889c85459 [Grammar Fix] GLM-4-MOE self.first_k_dense_replace is undefined. (#12455) Yuxuan Zhang 2025-11-06 00:03:45 +08:00
• 4d84f886e7 Refactor --debug-tensor-dump-layers to list (#12691) Yuhong Guo 2025-11-05 19:30:01 +08:00
• dc4f541823 fix trtllm_mla attention backend when disabling cuda graph. (#12687) yinghui 2025-11-05 01:35:02 -08:00
• 0648eb482d [Profiler] Add SGLANG_PROFILE_RECORD_SHAPES for recording shapes when profiling (#11641) zejunchen-zejunandHAI 2025-11-05 15:41:46 +08:00
• b88fab3111 fix: add seed bench_serving to cache key, remove redundant function definition. (#12680) yinghui 2025-11-04 23:39:11 -08:00
• 3694266051 Expand and update test coverage for AMD CI (#10044) Hubert Lu 2025-11-04 22:15:13 -08:00
• 9f5e701879 [router][grpc] Implement tool_choice support for Responses API (#12668) Chang Su 2025-11-04 21:43:42 -08:00
• cbf23dbbfa [Feature] add --lora-request-distribution arg to bench_serving.py and support skewed and distinct workloads (#12175) Glen Liu 2025-11-05 00:41:40 -05:00
• 6dade6c3b5 Fix VLLM dependency test (#12670) Kangyan-Zhou 2025-11-04 20:49:58 -08:00
• b419e20c5b [Dockerfile] Speed up docker image building (#8784) Yingchun Lai 2025-11-05 12:26:18 +08:00
• 48641435d6 fix typo of args description in sglang.profiler (#12486) 3 people 2025-11-05 12:15:13 +08:00
• 44b1b394a4 [PD-Disagg] Check finish after pop tranferred (#12638) Liangsheng Yin 2025-11-05 11:18:09 +08:00
• 0711d1509b [NVIDIA] Fix cutedsl backend of MoE (#12353) Kaixi Hou 2025-11-04 18:54:55 -08:00
• 09938e1f82 chore: bump SGLang version to 0.5.4.post3 (#12639) sglang-bot 2025-11-05 10:32:11 +08:00
• 2340798353 Register allgather/reducescatter buffers with symm memory (#12572) Nicolas Castet 2025-11-04 19:11:36 -06:00
• 1357ab025a [router][grpc] Emit OutputItemDone event and store output item array (#12656) Chang Su 2025-11-04 17:01:03 -08:00
• 44da737770 [fix] Handle escaped characters in GLM tool call parser to prevent double serialization (#12456) soaringk 2025-11-05 08:48:14 +08:00
• fb9582c4e1 Add multi-GPU configurations to nightly-test.yml (#12585) alisonshao 2025-11-04 16:46:30 -08:00
• d22d044734 Revert "Enable memory saver for hybrid model" (#12648) Baizhou Zhang 2025-11-04 16:22:06 -08:00
• 887742a1e7 [router][grpc] Fix index issues in reasoning content and missing streaming events (#12650) Chang Su 2025-11-04 15:38:43 -08:00
• 34f7564df0 [NVIDIA] Fix wrong symmetric sizes for fp4 cases (#12640) Kaixi Hou 2025-11-04 14:19:37 -08:00
• 1cfbbc42d8 [Bug] Fix NSA Backend KV-Buffer Shape Mismatch in DeepSeek-V3.2 (#12645) Johnsonms 2025-11-04 13:57:32 -08:00