Commit Graph

Select branches

Hide Pull Requests

bot/update-est-time-20260928

dsv41-pd

dsv41-pd-visioncp

main

Mono Color

• df9c831ab8 Unify think_end_id to model_config as single source of truth (#22148) Liangsheng Yin 2026-04-05 03:35:38 -07:00
• aeff9fb7c1 Add dump_metric to MMMU, lm-eval, and NeMo Skills eval paths (#22147) Liangsheng Yin 2026-04-05 03:23:52 -07:00
• cd2d45e220 Isolate spec V1 path in decode post-processing (#22146) Liangsheng Yin 2026-04-05 03:16:56 -07:00
• 106baedbfb [Doc] Update GLM-5 instructions in sglang documentation (#21716) Baizhou Zhang 2026-04-05 03:13:07 -07:00
• 10b18b8b29 [diffusion] Add is_float64_supported to Platform (#22112) R0CKSTAR 2026-04-05 18:12:28 +08:00
• 5a35316417 Enable IndexCache for DeepSeek V3.2 (#21405) iLeGendandBaizhou Zhang 2026-04-05 17:45:58 +08:00
• 088203454b [Fix] Fix nightly tests (#22140) Baizhou Zhang 2026-04-05 02:26:42 -07:00
• bd6a585605 Consolidate reasoning tests into test/registered/reasoning/ (#22139) Liangsheng Yin 2026-04-05 01:09:11 -07:00
• 2b119ba388 [diffusion] fix: fix accuracy for flux series (#22059) Yuhao YangandMick 2026-04-05 16:03:17 +08:00
• 71544f0341 [model] support voxtral (speech-to-text) (#21635) Xiancheng Mengandmengxiancheng03 2026-04-05 15:46:30 +08:00
• 904bb476d8 Migrate reasoning_tokens tests to existing server fixtures (#22102) Liangsheng Yin 2026-04-05 00:30:57 -07:00
• bb9e058f5b Add failfast flag to rerun-test workflow (#22141) Liangsheng Yin 2026-04-05 00:24:50 -07:00
• edee9ae929 Update ci_auto_bisect.py to use correct model (#22142) Kangyan-Zhou 2026-04-04 23:57:52 -07:00
• 723ed6c3c4 [CI]Temporary ban auto benchmark tool test (#22138) Baizhou Zhang 2026-04-04 23:18:19 -07:00
• 675100b53a Remove flaky TestToolChoiceLfm2Moe from test_tool_choice (#22137) Liangsheng Yin 2026-04-04 22:53:11 -07:00
• dd49127fe6 [AMD]: Support MLA with nhead<16 and FP8 KV cache for TP=8 (Kimi K2.5… (#21213) RoyWangandRoyWang 2026-04-05 13:13:29 +08:00
• 8cbeacd783 feat: CI auto-bisect workflow for automated regression analysis (#22119) Kangyan-ZhouandClaude Opus 4.6 2026-04-04 18:58:18 -07:00
• 70658bfeb5 fix: add missing f-string prefixes in warning and assert messages (#22067) 3 people 2026-04-05 08:20:16 +08:00
• efee62efa6 [diffusion] CI: improve diffusion comparison benchmark setting for realistic perf and auto-discover ut (#22086) MickandClaude Opus 4.6 2026-04-04 23:20:37 +08:00
• 0f0f004f1f [Benchmark] Add auto benchmark tool with YAML-driven server flag search and canonical dataset format (#21736) Xiaoyu Zhang 2026-04-04 21:46:58 +08:00
• da25b471e3 Align diffusion nightly presets and broaden skill discovery (#22099) Xiaoyu Zhang 2026-04-04 21:43:52 +08:00
• f3f7711dac Fix Python 3.11 f-string lint error in deepgemm Blackwell benchmark (#22108) Xiaoyu Zhang 2026-04-04 21:15:22 +08:00
• abc297521f Fix killall_sglang missing the main sglang serve process (#22103) Liangsheng Yin 2026-04-04 03:43:08 -07:00
• e9d92b0e33 Relax spec decoding accuracy threshold to fix flaky test (#22100) Liangsheng Yin 2026-04-04 02:38:35 -07:00
• 1ad6839659 [Feature] Add Reasoning Tokens Usage (#15562) 6 people 2026-04-04 17:18:10 +08:00
• bf984ae65d Revert "[Bugfix] Temporarily skip TRTLLM attention on (G)B300 (SM103) to avoid high-concurrency hang" (#22098) Baizhou Zhang 2026-04-04 02:17:19 -07:00
• 46bf19cdab chore: bump flashinfer version to 0.6.7.post2 (#22097) sglang-botandsglang-bot 2026-04-04 02:16:25 -07:00
• 24763256b9 [Speculative Decoding] Add FA4-based Spec Support (#21080) narutolhyandluhongyu.4869 2026-04-04 02:09:45 -07:00
• 34d5765e2f [VLM] Chunk-aware ViT encoding with per-image cache and lazy device transfer (#22038) Yuhao Yang 2026-04-04 16:55:17 +08:00
• b5e8c4b9e3 model: support LFM2-VL (Liquid Foundation Model 2 Vision-Language) (#21230) Piotr MazurekandPiotr Mazurek 2026-04-04 10:36:04 +02:00
• 1fb4bf3558 [diffusion] fix: validate attention backend for Ring Attention in USPAttention (#21828) R0CKSTAR 2026-04-04 16:24:02 +08:00
• 9fa12d605a Add dsv3 router gemm benchmark on blackwell (#17707) harrisonlimh 2026-04-04 01:18:01 -07:00
• 82ea4906cf [diffusion] Default NVFP4 to CUTLASS and add all-model shape benchmarks (#22091) Xiaoyu Zhang 2026-04-04 16:14:38 +08:00
• ff8e47edf9 [5/n] Lora support cuda graph (#21647) Ethan (Yusheng) Su 2026-04-04 00:31:46 -07:00
• a94c3804c2 fix: mistral embedding regression fix (#21913) Douglas Yang 2026-04-04 00:11:51 -07:00
• 005e582d06 [diffusion] improve: norm fusion for z-image (#18762) 3 people 2026-04-04 02:01:01 -04:00
• ef13031243 Tiny fix step3.5-flash launch crash (#22076) Qiaolin Yu 2026-04-03 22:25:25 -07:00
• 990c7590b8 [RL] Support mxfp8 DeepSeek V3 (#21280) Ziang Li 2026-04-03 21:57:45 -07:00
• 68f4c52d34 fix ut test_moe (#21735) Huaiyu, Zheng 2026-04-04 12:57:11 +08:00
• de9859073f Add --stream-response-default-include-usage server flag (#16711) faceless void 2026-04-04 12:36:00 +08:00
• 31c9d8e885 [Diffusion] Fix weight scale swizzle and add large-M kernel config for FLUX.2-dev-NVFP4 (#22064) CHEN Xi 2026-04-04 11:50:30 +08:00
• fe92f3563c dp: add profile req hook (#22083) Yilong Zhao 2026-04-03 20:47:09 -07:00
• b7ae3b5a9a GLM-4.7 and GLM-4.7-Flash Loading and import format (#21851) Yuxuan Zhang 2026-04-04 11:44:08 +08:00
• db3d4f4b76 [diffusion] model: support two stage pipeline of LTX-2 (#20707) 4 people 2026-04-04 09:37:28 +08:00
• 95cdbce34f [Test] Extract common PD server setup into base fixture (#22080) Liangsheng Yin 2026-04-03 16:37:12 -07:00
• 9593d434c4 fix: pause_generation should not populate running_batch on prefill nodes (#20273) Lawrence Wu 2026-04-03 16:16:06 -07:00
• 5118295f7b [CI] Support CPU stage and auto-batch same-stage files in /rerun-test (#22081) Liangsheng Yin 2026-04-03 15:56:54 -07:00
• 90e86800f4 [Score API] Implement EngineScoreMixin for scoring functionality and refactor Tok… (#21342) Sundara Raman Ramachandran 2026-04-03 15:17:42 -07:00
• ac1e437f6a Revert "[Feature] JIT activation and update skills (by codex)" (#22078) Baizhou Zhang 2026-04-03 15:04:15 -07:00
• 8cb337c8ea [Bugfix] Temporarily skip TRTLLM attention on (G)B300 (SM103) to avoid high-concurrency hang (#21906) Mohammad Miadh AngkadandBaizhou Zhang 2026-04-04 05:19:13 +08:00
• 1d7a53dd03 [Fix] XGrammarGrammarBackend reset to clear inherited cache (#22054) Yz Xiao 2026-04-04 05:17:59 +08:00
• 84118acf50 chore: bump sglang-kernel version to 0.4.1 (#22009) sglang-botandsglang-bot 2026-04-03 13:58:35 -07:00
• eb407b80f3 [Kernel] Make FA3/FA4 imports lazy in FlashAttentionBackend (#22028) Lianmin Zheng 2026-04-03 13:49:00 -07:00
• 6aafe756b9 Revert "[Feature] NVFP4 Marlin fallback for non-Blackwell GPUs (SM75+… (#22047) Brayden Zhong 2026-04-03 16:12:30 -04:00
• 0c9dc098e7 Fix DP attention worker port binding for IPv6 support (#21917) Shiyan Deng 2026-04-03 12:39:39 -07:00
• ed3435e37f [HiSparse]: Optimize server args checking-HiSparse is temporarily only available for DSA models. (#22065) Zhangheng 2026-04-04 02:23:56 +08:00
• 151f727163 [diffusion] fix: fix gated repo failing the generate cmd (#22040) 3 people 2026-04-04 00:43:11 +08:00
• 896ea75820 Remove reverted test (#22058) Ke Bao 2026-04-03 23:51:37 +08:00
• 47f4fd275a [CI] Fix test suite names and add suite validation (#21937) Ke Bao 2026-04-03 23:47:17 +08:00
• 44e5d35703 [Feature][JIT Kernel] JIT activation and update skills (by codex) (#21766) DarkSharpnessandweiminc 2026-04-03 23:28:54 +08:00
• 030fb1c4b1 refactor: replace mm_inputs dict with MultimodalProcessorOutput (#21738) Mick 2026-04-03 23:26:37 +08:00
• 9f409d0749 [CI] Adjust CI server launch timeout (#22045) Ke Bao 2026-04-03 22:38:07 +08:00
• ee9d922f5a Revert "[Kernel] Fuse temperature + softmax in sampling for decode speedup" (#22046) Xiaoyu Zhang 2026-04-03 21:32:08 +08:00
• 97adf8a290 [misc] Add hint for kernel release trigger (#22036) Baizhou Zhang 2026-04-03 03:31:44 -07:00
• 98ac40192b [Workflow] Fix kernel release build failures for aarch64 and wheel renaming (#22018) Baizhou ZhangandClaude Opus 4.6 2026-04-03 03:23:03 -07:00
• 838f815e9f [diffusion] CI: temporarily disable accuracy ci (#22031) Mick 2026-04-03 17:39:29 +08:00
• 56ac9c9932 [Fix] Add _MOE_TP to graph_capture for MoE models with ep>1 (#21907) 3 people 2026-04-03 02:33:16 -07:00
• ac593fed90 [AMD][Dockerfile] Support build-arg AITER_COMMIT for rocm.Dockerfile (#21949) Duyi-Wang 2026-04-03 16:54:28 +08:00
• cd75d54fc5 [Bugfix] Fix CUDA graph replay issues in trtllm_mla draft_extend (#21987) Khoa PhamandClaude Opus 4.6 2026-04-03 01:45:13 -07:00
• 4f84ce5807 [CI] ci: add test_http_server_auth.py to CI (#21866) shuwenn 2026-04-03 16:32:18 +08:00
• 658a2813d8 [NPU] Update CI Dependency (#21578) monkeyLoveding 2026-04-03 16:22:11 +08:00
• d07d0a15ce [AMD] Add MiniMax-M2.5 nightly perf benchmarks for MI30x and MI35x (#21524) Michael 2026-04-03 01:01:03 -07:00
• 7431db7392 [AMD] Enable FP8 KV cache and FP8 attention kernel for NSA on MI300/MI355 with TileLang backend (#21511) Thomas Wang 2026-04-03 15:58:23 +08:00
• ad0516d9c1 [NPU] optimize glm4.7 (#19246) Kelon 2026-04-03 15:44:07 +08:00
• d82097a0df [PD] Tiny register info field cleanup for mooncake backend (#22016) Shangming Cai 2026-04-03 15:13:44 +08:00
• 24f52e66d3 fix: remove duplicate words in comments (#22007) Ricardo-M-L 2026-04-03 15:05:39 +08:00
• 4cc970290d [CI] Fix duplicate job names that bypass branch protection (#22001) Liangsheng Yin 2026-04-02 23:59:35 -07:00
• 6b876a7710 [ROCM][RL] Shuffle Weight In-Place to Preserve Parameter Attributes (#21825) Yuzhen Zhou 2026-04-02 23:43:55 -07:00
• 75de479680 [Misc] Update CI permission (#22014) Baizhou Zhang 2026-04-02 23:37:05 -07:00
• 4d097047f2 [PD]: Add support for HiSparse to directly transfer the cache from Prefill to Decode DRAM. (#21591) 4 people 2026-04-03 14:06:12 +08:00
• 5c082c307a [Workflow] Fix kernel release jobs skipped on push events (#22011) Baizhou ZhangandClaude Opus 4.6 2026-04-02 23:03:13 -07:00
• 0a709cfe02 [Workflow] Avoid triggering nightly tests in kernel bump workflow (#22010) Baizhou Zhang 2026-04-02 22:40:33 -07:00
• 2c4fb88929 chore: bump sgl-kernel version to 0.4.1 (#21447) sglang-botandsglang-bot 2026-04-02 22:31:59 -07:00
• 5bcbc9757c [AMD] Resolve the performance degression when launch server with "--enable-aiter-allreduce-fusion" (#21947) kkandwunhuang 2026-04-03 13:10:24 +08:00
• d1b7c3907d [Parallel State Refactor 2/n] Unify code path of AMD deterministic all reduce (#20871) DarkSharpness 2026-04-03 12:33:17 +08:00
• 81efcc353a [NPU] Optimized the wording in the npu docs (#21998) amote-i 2026-04-03 11:51:40 +08:00
• efa7b2d5d3 Revert "[MUSA][9/N] Add FA3 attention backend support through MATE (MUSA AI Tensor Engine)" (#22002) Baizhou Zhang 2026-04-02 20:42:13 -07:00
• 5f0df1e2ad [Bugfix] Fix incorrect dp-attention parallel info in bench_one_batch (#21519) lviy 2026-04-03 11:13:53 +08:00
• 69e89a1fcc [VLM] Enable per-image MM splitting by default and remove MULTI_IMAGES modality (#21899) Yuhao Yang 2026-04-03 11:04:41 +08:00
• 8897ac58f0 [PP] qwen3 vl skip layer id for pp (#19135) narutolhy 2026-04-02 19:51:53 -07:00
• 991f3aa5b3 [Feature] NVFP4 Marlin fallback for non-Blackwell GPUs (SM75+) (#19652) Mook 2026-04-02 19:48:15 -07:00
• 2b5aed94f5 Remove maxItems=1 restriction when tool_choice is specified (#20208) Khoa Pham 2026-04-02 19:35:24 -07:00
• 0539c62bc1 [Diffusion][NPU] Add support for MOVA (#21633) Thomasandzhangshuai 2026-04-03 10:33:14 +08:00
• 1f97714f9b [CI] Add timeouts to Slack upload urlopen and WebClient (#21903) Kangyan-ZhouandClaude Opus 4.6 2026-04-02 19:30:55 -07:00
• 89affff290 Skip broken AutoModel mapping entries when resolving Llava submodules (#21892) Xiaoyu Zhang 2026-04-03 09:04:26 +08:00
• 29d8e959d7 [CI] Remove stale Ascend suite entries from test/srt/run_suite.py (#21978) Baizhou ZhangandClaude Opus 4.6 2026-04-02 16:47:19 -07:00
• 34ddf135fd [Feature] Stronger transformers modeling backend with TP, PP, MoE, VLMs, and torch compile (#19163) Adarsh Shirawalmath 2026-04-03 04:32:33 +05:30
• 939cf398a9 [MUSA][9/N] Add FA3 attention backend support through MATE (MUSA AI Tensor Engine) (#17985) oriandR0CKSTAR 2026-04-03 06:04:31 +08:00
• 566b4a4f1c [4/n] Support gpt oss 20b lora (#21570) Ethan (Yusheng) Su 2026-04-02 12:57:38 -07:00
• fe38410c3e Remove logging for subprocess watchdog start (#21968) Lianmin Zheng 2026-04-02 11:30:33 -07:00