chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity
9,282 Commits 1 Branch 0 Tags
e9d727cb9218f30a7352f2d17614d74ec0577062
Commit Graph

1703 Commits

Author SHA1 Message Date
Артем Савкин
b77b0ffd60 [NPU] NZ for non-quantized MOE, Qwen3 MOE double memory consumption fix (#15904) 
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
 2026-01-29 00:55:08 +08:00
Jinn
1953efb60e [AMD] ROCm: route W4A16 MoE to Triton and fix packed-weight loading (#17863) 2026-01-28 08:20:23 -08:00
Ziang Li
a8dda2aa57 [DSv32] Overlap indexer qk projection and activation quant (#17688) 2026-01-28 11:46:49 +08:00
Minglei Zhu
d90c0837e5 [hybrid-model] clean up and consolidate redundant fields in RadixLinearAttention (#17660) 2026-01-27 10:37:58 -08:00
monkeyLoveding
d578b41bad [NPU] Adapt cann 8.5: use sfa and lightning indexer op from cann and CI update (#17615) 
Co-authored-by: Kelon <kelonlu@163.com>
 2026-01-27 19:03:53 +08:00
MikkoParkkola
c56d19b977 fix(quantization): add sgl_kernel fallback for FP4 quantize on Blackwell GPUs (#17816) 2026-01-27 18:43:17 +08:00
Yuxuan Zhang
7106f6c8e1 [GLM-OCR] Support GLM-OCR Model (#17582) 
Signed-off-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
Co-authored-by: Xinyuan Tong <115166877+JustinTong0323@users.noreply.github.com>
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
 2026-01-26 22:24:00 -08:00
Yuhao Yang
479ab7a4e7 model: support Kimi-K2.5 (#17789) 
Co-authored-by: Mick <mickjagger19@icloud.com>
 2026-01-27 10:57:00 +08:00
lawtherWu
b56366f827 [NPU]DeepSeek-V3.2 support npu mlaprolog (#15381) 
Co-authored-by: Zhengda Qin <zhengdqin@gmail.com>
Co-authored-by: richhuan <huan_rz@qq.com>
 2026-01-26 20:42:37 +08:00
CSWYF3634076
1a19b3987d [Model] Add Ernie4.5 VL model support (#15679) 
Signed-off-by: CSWYF3634076 <wangyafeng@baidu.com>
Signed-off-by: wangyafeng <wangyafeng@baidu.com>
 2026-01-25 22:36:29 -08:00
HandH1998
a883906a24 Support mxint4 flashinfer_trtllm moe gemm (#16892) 2026-01-26 00:15:53 +08:00
Mohammad Miadh Angkad
1674b9ef44 [DeepSeek-V3.2] Fix TRT-LLM NSA in target_verify/draft_extend (#17662) 2026-01-25 13:10:14 +08:00
Trevor Morris
2c2c4e446b [NVIDIA] Add flashinfer all-to-all MOE dispatcher (#14668) 2026-01-24 22:59:55 +08:00
Yuan Luo
0c8165ffbd [Kimi-Linear] Refactor Kimi-Linear to support RadixLinearAttention (#17506) 
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
 2026-01-24 21:27:13 +08:00
Xiaoyu Zhang
3992a023e6 Move fa4 from sgl-kernel to jit kernel (#17353) 2026-01-24 15:25:03 +08:00
Lianmin Zheng
bc6f0b5ce7 [Auto Sync] Update logits_processor.py, test_logprobs.py (20260124) (#17664) 
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: yehu-ux <yehu@x.ai>
 2026-01-23 17:57:41 -08:00
McZyWu
8a5ed2434f [NPU]support model MiniCPM3-4B for npu (#16866) 2026-01-24 08:25:12 +08:00
Nan Jiang
ad05782160 fix post_residual_addition more generally (#17286) 2026-01-23 15:43:37 -08:00
Yi Zhong
08fcda2f63 add the fa4 mm backend and varlen func (#13539) 
Signed-off-by: vincentzed <207368749+vincentzed@users.noreply.github.com>
Co-authored-by: Brayden Zhong <b8zhong@uwaterloo.ca>
 2026-01-23 23:12:06 +08:00
akhilg-nv
2fb328109f [DeepSeek V3.2] Enable trtllm NSA with bf16 kvcache (#16758) 
Co-authored-by: DarkSharpness <76582120+DarkSharpness@users.noreply.github.com>
 2026-01-23 20:26:21 +08:00
Lianmin Zheng
56e6652d1d Lazy import torchao (#17626) 2026-01-22 22:04:51 -08:00
wufann
a921029b97 [AMD] Support ds3.2 on gfx942 platform (#17504) 
Co-authored-by: Hubert Lu <55214931+hubertlu-tw@users.noreply.github.com>
 2026-01-22 13:57:08 -08:00
Baizhou Zhang
283a2daeaa [hotfix] Reenable all reduce fusion on sm100 (#17591) 2026-01-22 23:36:38 +08:00
chenxu214
5d299c25c0 [NPU] bugfix with Kimi-k2 and bge-reranker-v2 model (#17478) 
Co-authored-by: amote-i <49533125+amote-i@users.noreply.github.com>
Co-authored-by: cy <chenyang08056032@163.com>
 2026-01-22 22:02:05 +08:00
Minglei Zhu
419bbcee10 refactor Qwen3-Next with a new RadixLinearAttention (#17373) 2026-01-22 17:42:06 +08:00
Baizhou Zhang
e2d33531f3 [Kernel] Little refactor of flashinfer allreduce norm fusion (#17474) 2026-01-22 13:31:57 +08:00
YC Tseng
17807caf82 [AMD] fix amd ci dpskv32 (#17432) 
Co-authored-by: michaelzhang-ai <michaelzhang.ai@users.noreply.github.com>
 2026-01-21 20:34:24 -08:00
Baizhou Zhang
3373545b9f [HotFix]Fix dtype mismatch in nsa indexer on AMD device (#17518) 2026-01-22 12:17:02 +08:00
Lianmin Zheng
1fdf5cac39 [Auto Sync] Update environ.py, fp8.py (20260121) (#17486) 
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Binyao Jiang <byjiang1996@gmail.com>
 2026-01-21 12:04:09 -08:00
Qiaolin Yu
4f6f5d25c8 Support fa4 decoding (#16034) 2026-01-21 22:54:02 +08:00
strgrb
bcc6d84f93 Use fused_sigmoid_gating_delta_rule_update_kernel for KDA (#17108) 2026-01-21 19:24:29 +08:00
Sam Shleifer
0d49b13fdd Fix circular import in quantization modules (#17372) 2026-01-21 15:47:09 +08:00
Baizhou Zhang
8251a74d5f [Tiny] Backward compatibility for fp4 gemm flags (#17466) 2026-01-21 14:34:40 +08:00
Ke Bao
d97066d209 Use attn_tp_group for all reduce in token embedding (#17403) 2026-01-20 23:56:21 +08:00
BingjiaWang
612026ad2c [OPT] DeepSeekV3.2: optimize indexer weight_proj-mma performance (#17205) 2026-01-20 23:09:44 +08:00
Yuan Luo
e6b7c04947 [Kimi-Linear] Refactor kimi-linear gate calculation to avoid duplicated code (#17160) 
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
 2026-01-20 14:29:24 +08:00
Thomas Wang
6988a0f570 Disable mla persistent kernel when not using fp8 kv_cache (#17327) 2026-01-19 20:29:54 -08:00
shuwenn
8fb45523f3 feat: support bitsandbytes quantization algorithm (#15325) 2026-01-19 18:36:56 -08:00
b8zhong
f374623fa9 [Refactor] Set fp4-gemm-backend=auto on SM100 and rename fp4-gemm-backend with flashinfer_ prefix (#17309) 2026-01-19 20:09:07 +08:00
Shu Wang
5c02217746 Inclusion of nvfp4 blockscale in EPLB Rebalance (#17158) 2026-01-19 17:45:27 +08:00
yudian0504
9fe56cd0fb Fix kernel selection in biased_grouped_topk_gpu (#17325) 2026-01-19 14:06:11 +08:00
Yongfei Xu
d2105d4abd [DeepSeek v3.2] Opt MTP decode cuda batch sizes and nsa implementation (#16961) 2026-01-19 11:54:11 +08:00
Lee Nau
84c8390514 Use dsv3 optimized routing fused_topk_deepseek instead of moe_fused_gate (#15347) 2026-01-19 11:50:16 +08:00
Koushik Dutta
d3eafc7357 [GLM 4.7] Add RTX 6000 Pro aka sm120 (#17235) 
Co-authored-by: root <root@ubuntu-nvidia.localdomain>
 2026-01-18 13:33:19 -08:00
Jinyan Chen
e00b43442d [jit-kernel] Add CuTe DSL GDN Decode Kernel (#15631) 
Co-authored-by: Jinyan Chen <jinyanc@nvidia.com>
 2026-01-18 12:54:36 -08:00
Todobe
733de6be31 [NPU]Support GPT-OSS for NPU (#14197) 2026-01-19 04:13:41 +08:00
b8zhong
4df74eb576 [Refactor] Add -fp4-gemm-backend to replace SGLANG_FLASHINFER_FP4_GEMM_BACKEND (#16534) 
Co-authored-by: Vincent Zhong <207368749+vincentzed@users.noreply.github.com>
 2026-01-18 23:25:46 +08:00
Nan Jiang
dd99f818e0 fix: fix regression and unclear pattern (#16561) 2026-01-16 23:21:42 -08:00
b8zhong
d36f6f043c [Fix] flashinfer_trtllm intermediate_size assertion with Qwen3 + TP=8 (#16824) 2026-01-17 09:24:05 +08:00
Baizhou Zhang
a04675892e Update flashinfer to 0.6.1 (#15551) 2026-01-17 00:48:30 +08:00