kvcache-ai/sglang
1
0
Fork 0
mirror of https://github.com/kvcache-ai/sglang.git synced 2026-07-08 16:27:09 +00:00
Code Issues Packages Projects Releases Wiki Activity
8,783 Commits 157 Branches 0 Tags
675acecec63f439d515881707e04ba8faeccbbcd
Commit Graph

5674 Commits

Author SHA1 Message Date
Zhiqiang Xie
675acecec6 Attention backend selection bug fix for hicache (#16779) 
Co-authored-by: hxie <haiyong.xie@ustc.edu>
 2026-01-10 00:22:41 -08:00
Insideyyy
67b61a4e8d [Rework] Add SwapAB Optimization for triton fused_moe_kernel on SM90. (#16723) 2026-01-10 13:57:44 +08:00
Shifang Xu
d27f16f38a Fix EPLB + FP4 Quantization Compatibility Issue (#13715) 
Co-authored-by: Shu Wang <shuw@nvidia.com>
 2026-01-10 13:38:19 +08:00
fzyzcjy
c89949bbaf Tiny let soft watchdog cover initialization phase (#16853) 2026-01-10 13:20:08 +08:00
Ziang Li
20abaee26c [DSv32] Overlap indexer weights_proj during dual_stream decode (#16637) 
Co-authored-by: Ziang Li <ziangli@humansand.ai>
 2026-01-10 13:06:44 +08:00
fzyzcjy
32a569fb77 Tiny add CPU resource monitoring for overload diagnosis (#16852) 2026-01-10 12:48:18 +08:00
fzyzcjy
3ed3b7ef7c Tiny add routing key distribution metrics (#16847) 2026-01-10 12:06:40 +08:00
fzyzcjy
1f9d4795a9 Tiny add gauge histogram abstraction for engine and router (#16848) 2026-01-10 11:45:25 +08:00
Liangsheng Yin
e6d40bff81 Revert "feat: reduce constrained-decoding overhead in TP" (#16845) 2026-01-10 11:39:38 +08:00
若可
fbc128a32e fix(function_call): group batch decode by options instead of fallback (#16698) 
Co-authored-by: Muqi Li <muqi1029@gmail.com>
 2026-01-10 11:38:13 +08:00
Yinghai Lu
e91a717632 [llama] Allow passing tp_rank and tp_size into llama mlp (#16837) 2026-01-10 10:32:05 +08:00
fzyzcjy
1f0ea4f958 Add routing key based schedule policy (#16840) 2026-01-10 10:16:23 +08:00
fzyzcjy
15da3061bc Tiny pass routing key to scheduler processes (#16839) 2026-01-10 10:13:00 +08:00
fzyzcjy
6406a5969b Tiny support customizing prometheus buckets for prefill delayer (#16831) 2026-01-10 08:17:37 +08:00
fzyzcjy
cec19b56c6 Tiny add command line args for prefill delayer and unify names (#16830) 2026-01-10 07:54:25 +08:00
Yi Zhong
84ab32a2a0 Reduce some small cpu overhead in stream fetch (#16587) 
Signed-off-by: vincentzed <207368749+vincentzed@users.noreply.github.com>
 2026-01-09 21:53:06 +00:00
Leoyzen
8ef5b90528 Fix GLM-4.7 MoE Detector complex JSON Schema type parsing (#15753) 2026-01-10 01:21:10 +08:00
Raayan Dhar
76b3c698d6 feat: reduce constrained-decoding overhead in TP (#13947) 
Signed-off-by: Raayan Dhar raayan.dhar@gmail.com <raayan.dhar@gmail.com>
Signed-off-by: raayandhar <raayan.dhar@gmail.com>
 2026-01-10 00:38:32 +08:00
Hanming Lu
5dcff94791 [nemetron/mtp] fix nemotron mtp (#16275) 2026-01-09 23:22:21 +08:00
fzyzcjy
8eeffbe9aa Support token low usage watermark in prefill delayer (#16814) 2026-01-09 23:07:18 +08:00
fzyzcjy
7656d26737 Tiny enhance prefill delayer observability (#16812) 2026-01-09 22:51:16 +08:00
fzyzcjy
dd24ba9084 Refactor prefill delayer for clarity and extensibility (#16811) 2026-01-09 22:46:41 +08:00
siyu
068abe7e40 add doc for #14386 (#14655) 2026-01-09 22:38:51 +08:00
sunxxuns
64a31d4b75 [diffusion] amd: fix SGLANG_DIFFUSION_ATTENTION_BACKEND env var for diffusion attention backend selection (#16325) 
Co-authored-by: root <root@mi300x8-008.atl1.do.cpe.ice.amd.com>
 2026-01-09 20:16:58 +08:00
ybyang
2babf88f24 fix spec qwen3 pd error (#16708) 2026-01-09 19:20:17 +08:00
Shangming Cai
0c4e155a3c chore: bump mooncake version to 0.3.8.post1 (#16792) 
Signed-off-by: Shangming Cai <csmthu@gmail.com>
 2026-01-09 18:42:27 +08:00
Hubert Lu
d6d5c3fdea [AMD] Clean up vllm dependencies in moe_runner/triton.py (#11349) 
Co-authored-by: HAI <hixiao@gmail.com>
 2026-01-09 00:24:04 -08:00
Alison Shao
e46f79431b Fix external_models import path and migrate model loading tests (#16458) 2026-01-08 23:43:49 -08:00
Xiaoyu Zhang
9d4d57dbfa [Diffusion] Tiny rename parallel_groups (#16743) 
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
 2026-01-09 15:08:25 +08:00
Hanming Lu
41609b52fe support page size large than 64 for mamba radix cache with fix (#16768) 2026-01-09 13:28:37 +08:00
Fenglin Yu
c1c13c841a [diffusion] docs: add LoRA support (#16378) 2026-01-09 11:31:38 +08:00
Liangsheng Yin
75da784d48 Tiny simplify draft worker init. (#16446) 2026-01-09 11:20:29 +08:00
Yongfei Xu
05dfef92a1 [DeepSeek 3.2] Support and optimize pipeline parallelis when context pipeline enabled (#16380) 
Co-authored-by: ybyang <ybyang7@iflytek.com>
 2026-01-09 11:01:49 +08:00
triple-mu
f7f5c3896d [diffusion] model: wan tp+usp optimize (#16720) 
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Xiaoyu Zhang <35585791+BBuf@users.noreply.github.com>
 2026-01-09 09:40:31 +08:00
Simo Lin
55a8dd0095 [grpc] Fix protobuf compilation in isolated build environments (#16754) 2026-01-08 13:09:01 -08:00
Chang Su
16880235d1 [grpc] Auto-generate protobuf files during wheel build (#16409) 2026-01-08 09:09:54 -08:00
Yi Zhong
cda35611d4 [perf] Add two stream norm for Olmo3 speedup 5% (#13681) 
Signed-off-by: Vincent Zhong <207368749+vincentzed@users.noreply.github.com>
Co-authored-by: b8zhong <b8zhong@uwaterloo.ca>
 2026-01-08 17:01:35 +00:00
Mohammad Miadh Angkad
05ab110e02 [Performance] Force split_k=1 for MXFP4 Triton kernels on Hopper (#16014) 
Signed-off-by: Mohammad Miadh Angkad <MAngkad.BSDSBA2027@aim.edu>
 2026-01-08 23:52:22 +08:00
cheng peng
82a8d77bc0 [Diffusion] model: fix zimage tp (#16719) 
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
 2026-01-08 22:47:21 +08:00
Xiaoyu Zhang
294ff71d18 [Diffusion] Avoid cpu2gpu sync in flashinfer rope and apply flashinfer rope to wanvideo (#16668) 
Co-authored-by: Mick <mickjagger19@icloud.com>
 2026-01-08 22:44:38 +08:00
fzyzcjy
b6e8a0d851 Add file size hints in bitwise model file verifier (#16735) 2026-01-08 22:28:52 +08:00
luoyuyan
fb7609f1dd Fix FP8 MoE NaN with DeepGEMM on Blackwell (#16622) 2026-01-08 22:24:12 +08:00
yuhao
d2ea44f775 VLM: enhance VL embedding model with video input support and revise warm-up strategy (#16635) 
Co-authored-by: Mick <mickjagger19@icloud.com>
 2026-01-08 22:12:01 +08:00
fzyzcjy
83abecd0c2 Support pre-generating and using expected checksums (#16730) 2026-01-08 20:26:13 +08:00
fzyzcjy
d54f0a10b4 Support bitwise weight checksum verifier (#16729) 2026-01-08 20:19:25 +08:00
Lianmin Zheng
fb04e7e3c8 [Auto Sync] Update schedule_batch.py, common.py, eagle_info... (20260105) (#16519) 
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: cctry <shiyang@x.ai>
 2026-01-08 02:35:46 -08:00
chenxu214
7dd679cbb9 [NPU][Bugfix] Fix qwen3 error when enable-dp-lm-head (#16115) 2026-01-08 15:15:43 +08:00
陈一涵
48b8dcd42e [jit kernel] support dtype as a cpp template parameter (#16452) 2026-01-08 13:54:33 +08:00
Fan Lin
41b434a7e6 [diffusion] endpoint: add API endpoint to query loaded LoRA adapters information (#16533) 2026-01-08 13:51:26 +08:00
Hubert Lu
4935344fcd [AMD] Fix aiter page-size handling, DeepSeek MLA tuple inputs, and HiCache/FA3 decode-backend override (#16531) 2026-01-07 21:14:32 -08:00