chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 11 Packages Projects Releases Wiki Activity
10,502 Commits 1 Branch 0 Tags
ed42af99a92fa6f69c5de7f05d023b5a673ddf6b
Commit Graph

1934 Commits

Author SHA1 Message Date
Артем Савкин
ed42af99a9 [NPU] [Quantization] w4a4 MoE layer support (#18924) 2026-03-11 16:52:35 +03:00
Yoray Zack
9991debde3 [Feature] Integrate Elastic NIXL-EP into SGLang (#19248) 
Signed-off-by: Barak Biber <bbiber@nvidia.com>
Signed-off-by: Yoray Zack <yorayz@nvidia.com>
Signed-off-by: Itay Alroy <ialroy@nvidia.com>
Co-authored-by: Barak Biber <bbiber@nvidia.com>
 2026-03-11 17:37:43 +08:00
xieminghe1
21a0015aa3 [PCG]add piecewise cuda graph support for marlin linear (#20119) 
Co-authored-by: undefined <zhouchen.arrebol@jd.com>
 2026-03-11 10:57:08 +08:00
Kurkur
16ec4f3a4a Integrate the AddRmsNorm operator (#19939) 2026-03-11 09:05:04 +08:00
Qiaolin Yu
09a118fafe Support return_logprob for spec v2 (overlap safe) (#19801) 
Co-authored-by: Ratish1 <ratish1501@gmail.com>
Co-authored-by: Ratish1 <formula733@gmail.com>
Co-authored-by: hnyls2002 <lsyincs@gmail.com>
 2026-03-10 15:38:27 -07:00
Ziang Li
76ee4bb98c [FlashInfer v0.6.4] [RL] Integrate FlashInfer mxfp8 gemm, MoE, and routed MoE (#19537) 2026-03-10 15:37:57 -07:00
R0CKSTAR
db97f193b7 [diffusion][llm] macOS support (#19549) 
Signed-off-by: Xiaodong Ye <yeahdongcn@gmail.com>
Co-authored-by: Mick <mickjagger19@icloud.com>
 2026-03-10 13:11:07 -07:00
Hexq0210
9884957c07 [NPU] Bugfix for qwen35 on NPU (#19756) 2026-03-10 20:03:26 +08:00
Xiaoyu Zhang
51d9d34977 [2/n jit_kernel restruct] unify rotary embedding entrypoints under rope.py (#20247) 2026-03-10 17:49:57 +08:00
Thomas Wang
6407891b4f [AMD] Fp8 prefill integration with radix cache path for dpsk models (#20187) 2026-03-10 02:49:47 -07:00
Xiaoyu Zhang
c812504b92 [1/n jit_kernel restruct] unify cache usage and clean up naming in ngram_embedding (#20244) 2026-03-10 15:53:43 +08:00
khalilzhk
5f717913a0 support Kimi-K2.5-w4a8 on ascend 2026-03-10 14:43:27 +08:00
Baizhou Zhang
be63f982b7 [V32/GLM5] Control the threshold of applying dense attention with an environ (#20062) 2026-03-09 14:36:10 -07:00
Martin Vit
d39ed074cf fix: default FP4 GEMM backend to flashinfer_cudnn on SM120 (Blackwell) (#20047) 
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Brayden Zhong <b8zhong@uwaterloo.ca>
 2026-03-09 14:13:08 -07:00
AMD-yanfeiwang
f0153ad225 [AMD][Feature] support fp4 dispatch and fp8 combine in moriep (#19757) 
Co-authored-by: Duyi-Wang <duyi.wang@amd.com>
 2026-03-09 12:52:05 -07:00
sjqgogogogo
eb4ba1bde2 Feature/support longcat flash lite (#17838) 
Co-authored-by: sunjiaqi11 <sunjiaqi11@meituan.com>
Co-authored-by: ispobock <ispobaoke@gmail.com>
 2026-03-09 23:00:11 +08:00
roikoren755
c76251f70c Return intermediate Mamba states (#19716) 2026-03-09 16:04:36 +08:00
cen121212
fc543df289 [NPU] qwen3_vl encoder support graph 2026-03-09 10:13:35 +08:00
Bingxu Chen
17721b00fd [AMD] Fix Tensor Memory Aliasing (#19928) 2026-03-07 08:06:10 -08:00
YeChang Guo
13bdc7bf4a [Feature][NPU]: add runtime support for AutoRound quantized models (#16699) 
Co-authored-by: root <root@localhost.localdomain>
Co-authored-by: ronnie_zheng <zl19940307@163.com>
 2026-03-07 18:03:55 +03:00
Fan Yin
43d6a32045 [sgl-kernel] rebase FlashMLA 0217 (#18902) 
Co-authored-by: Baizhou Zhang <sobereddiezhang@gmail.com>
 2026-03-07 00:30:52 -08:00
Qiaolin Yu
925185f9ec Fix flashinfer backend with pcg (#20061) 2026-03-06 20:01:43 -08:00
Mohammad Miadh Angkad
f88acf8780 [JIT Kernel] Reland NVFP4 kernels to JIT (#20012) 2026-03-07 10:31:08 +08:00
Shu Wang
61de303f0a Fix fallback to default tactic (flashinfer autotuner) with trtllm_fp4_block_scale_moe (#19189) 2026-03-06 15:15:04 -08:00
Mohammad Miadh Angkad
759700c808 Fix SM120 triton_kernels MXFP4 block_k for GPT-OSS (#20040) 2026-03-06 10:53:08 -08:00
R0CKSTAR
de1a0afcbc [MUSA][10/N] Add GGUF support (#18357) 
Signed-off-by: Xiaodong Ye <xiaodong.ye@mthreads.com>
 2026-03-06 10:50:35 -08:00
Yuan Luo
f7de9375ac [GDN][Qwen3-Next][Qwen3.5] Fuse fused_gdn_gating and fused_recurrent_gated_delta_rule_update in verify_target (#19775) 2026-03-06 21:42:44 +08:00
Clint
27053aa5ed Fix MLA decode path returning unwritten (padded) rows (#19902) 2026-03-06 00:54:29 -08:00
Baizhou Zhang
51e5dc845a Revert "[Kernel Slimming] Migrate NVFP4 kernels to JIT" (#20005) 2026-03-05 19:40:00 -08:00
Xinyuan Tong
9ebffef1ef [FIX] NSA backend page_table overflow in speculative decoding target_verify (#19016) 2026-03-05 16:04:58 -08:00
Qiaolin Yu
46dced64ea Adjust padding size to improve triton_kernels moe performance (#19174) 2026-03-05 14:50:40 -08:00
akhilg-nv
73d272bddb Revised fix for HybridAttnBackend forward for linear attn (#19369) 2026-03-06 00:05:35 +08:00
StonyPort
806d41ab65 [quant] fix fp32 downcasting (#19844) 
Co-authored-by: qiuxuan.lzw <qiuxuan.lzw@alibaba-inc.com>
 2026-03-05 17:54:59 +08:00
Rain Jiang
472eef4071 fa4 cleanup (#19727) 2026-03-05 17:54:25 +08:00
Mohammad Miadh Angkad
2bdd89a6cd [Kernel Slimming] Migrate NVFP4 kernels to JIT (#19437) 2026-03-05 15:22:28 +08:00
Xinyi Song
0e6a64712a [bugfix] Fix PPMissingLayer AttributeError when Using PP (#19804) 2026-03-04 19:48:15 -08:00
kpham-sgl
376dfb03f7 Fix issue 19717 by making qo_indptr uniform strided instead of packed (#19807) 2026-03-04 15:27:10 -08:00
Bingxu Chen
329817e262 [AMD] Move get_global_server_args import out of CUDA-only block to fix NameError on AMD (#19866) 2026-03-04 10:23:42 -08:00
strgrb
738ebfd330 KDA: fuse qkv conv and support stride for fused_sigmoid_gating_delta_rule_update_kernel (#19506) 2026-03-04 22:45:53 +08:00
YeChang Guo
6910c1b281 [Feature][NPU]: add runtime support for GPTQ-quantized MoE models (#16364) 
Co-authored-by: GuoYechang <52730608+GuoYechang@users.noreply.github.com>
Co-authored-by: root <root@localhost.localdomain>
 2026-03-04 16:02:19 +03:00
chenxu214
88cfa6c11d [NPU]Releasing redundant memory of w13_weight and nz when the ascend_fuseep feature is enabled (#19813) 2026-03-04 19:26:29 +08:00
Michael
c6850ac30c [AMD] Fix Qwen3-Coder-Next: Add missing k_scale/v_scale args to extend_attention_fwd in aiter_backend (#19736) 2026-03-03 22:01:08 -08:00
sglang-bot
b7f7df7ee6 [NSA] Fix line-too-long lint in can_nsa_prefill_cp_round_robin_split (#19829) 2026-03-03 20:34:22 -08:00
Baidu-AIAK
6851613b93 [Bugfix] For cp: Fixed hang problem in prefix cache and kvcache support fp8 in-seq-split mode (#19656) 
Co-authored-by: vincent <vincent@vincentdeMacBook-Pro.local>
 2026-03-03 19:19:46 -08:00
xieminghe1
ee5ccde0ad support fused_moe_triton and moe_sum_all_reduce kernel fusion[reduce … (#19672) 
Co-authored-by: undefined <zhouchen.arrebol@jd.com>
 2026-03-04 10:30:33 +08:00
Brayden Zhong
e2af840c3d Various SM120 improvements (#19721) 2026-03-03 16:46:13 -08:00
Hubert Lu
441045a7bf [AMD] Fix EAGLE3 speculative decoding with aiter attention backend (#19362) 2026-03-03 16:12:13 -08:00
Yi Zhong
b8c71f895e Add tuned triton==3.5.1 h200 tp2, tp4 for qwen 3 next (#15948) 
Signed-off-by: vincentzed
 2026-03-03 14:47:19 -08:00
Yi Zhong
0c760c4cd7 Add tuned triton==3.5.1 b200 tp2, tp4 for qwen 3 next (#15917) 
Signed-off-by: vincentzed <207368749+vincentzed@users.noreply.github.com>
 2026-03-03 14:47:05 -08:00
Brayden Zhong
9305f0e58d Support triton_kernels for GPT-OSS on SM120 (#19718) 
Co-authored-by: amittell 1388680+amittell@users.noreply.github.com
 2026-03-03 14:14:01 -08:00