EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity
2,837 Commits 8 Branches 0 Tags
33b16ad178573e54eba44999c5af4740b57affd1
Commit Graph

1873 Commits

Author SHA1 Message Date
Liangsheng Yin
33b16ad178 Distinguish bootstrap key only in decode server (#5422) 2025-04-15 20:59:28 +08:00
shangmingc
ffde65a094 [PD] Fix dynamic port support and MLA buffer for Mooncake (#5415) 
Signed-off-by: Shangming Cai <caishangming@linux.alibaba.com>
Co-authored-by: ybyang <ybyang7@iflytek.com>
 2025-04-15 19:29:31 +08:00
lambert0312
471650dee0 Fix broadcast use cuda device lead to memory capacity unbalanced (#5416) 2025-04-15 02:47:26 -07:00
Yuan Luo
d06a83fb01 Support dynamic connection and TP 16 (#5351) 
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
 2025-04-15 17:08:07 +08:00
Zhaoyang Hao
5d13440162 [FIX] Fix concatenation error in capture_bs when open --disable-cuda-graph-padding and without MTP (#5412) 2025-04-15 01:42:27 -07:00
Yuhong Guo
3dfc6023ce Fix bench_serving with random-ids (#5214) 2025-04-15 01:34:35 -07:00
fzyzcjy
15e91d721b Tiny fix DeepseekScalingRotaryEmbedding always use forward_native (#5406) 2025-04-15 01:33:47 -07:00
Yineng Zhang
8aab7fdb21 chore: upgrade sgl-kernel 0.0.9 (#5401) 2025-04-14 22:37:59 -07:00
Yangcheng Li
ee9d6ca677 [fix/misc] remove duplicate row in deepseek v2 model (#5279) 2025-04-14 18:41:24 -07:00
Ximingwang-09
2dd6489468 Add H20 dtype fp8_w8a8 shared experts fused MoE kernel tuning configs for DeepSeek V3/R1 (#5291) 
Co-authored-by: ximing.wxm <ximing.wxm@antgroup.com>
 2025-04-14 18:40:31 -07:00
lambert0312
61e7c4dd21 Add A800 shared experts fused MoE kernel tuning configs for DeepSeek V3/R1 (#5368) 2025-04-14 18:39:44 -07:00
Baizhou Zhang
f6772f1497 [Fix] Turn off DeepGEMM by default (#5263) 2025-04-14 17:45:44 -07:00
Xiaoyu Zhang
38076dea84 apply fused moe gate in ds v3/r1 (#5371) 
Co-authored-by: Yineng Zhang <me@zhyncs.com>
 2025-04-14 16:24:26 -07:00
Ke Bao
5e0a9b0981 Apply deepseek cuda rope (#5385) 
Co-authored-by: Yineng Zhang <me@zhyncs.com>
 2025-04-14 15:22:43 -07:00
JieXin Liang
bdde237562 [perf] experimental enhance fp8 per-tensor quant (#5370) 2025-04-14 12:35:43 -07:00
ybyang
e9fc2ac7b6 [PD Bug] fix MLA get_contiguous_buf_infos error (#5384) 2025-04-14 22:56:39 +08:00
Liangsheng Yin
44afde82d7 Fix PD disaggregation bugs (#5326) 2025-04-14 19:27:30 +08:00
yhyang201
072df75354 Support for Qwen2.5-VL Model in bitsandbytes Format (#5003) 2025-04-14 02:03:40 -07:00
fzyzcjy
defede5073 Fix DeepSeek DP Attention + torch compile (#5367) 
Co-authored-by: ispobock <ispobaoke@163.com>
 2025-04-14 01:07:58 -07:00
Yongtong Wu
14e8bd889f Free metadata_buffer_index after transfer finished (#5364) 2025-04-14 16:04:46 +08:00
yulei
adca585bfb [DeepEP] Reduce routed scaling overhead (#5277) 
Co-authored-by: Cheng Wan <54331508+ch-wan@users.noreply.github.com>
 2025-04-13 16:03:09 -07:00
Yineng Zhang
39e411385c fix #5322 (#5359) 2025-04-13 13:57:36 -07:00
huangtingwei
5fbafbb8f8 fix MLATokenToKVPoolHost get_size_per_token bug (#5161) 
Co-authored-by: AniZpZ <zhuangsen.zp@antgroup.com>
 2025-04-13 12:37:26 -07:00
Byron Hsu
a9499885e9 [PD] Add transfer backend abstraction (#5328) 2025-04-14 01:39:39 +08:00
Liangsheng Yin
f765579046 Fix typo: infight -> inflight (#5357) 2025-04-14 01:25:30 +08:00
Yineng Zhang
f58b929a51 chore: upgrade sgl-kernel 0.0.8.post3 (#5342) 2025-04-13 00:45:59 -07:00
mlmz
8311b07fb9 Fix: Ensure tensors for dist.broadcast match NCCL backend device (#5322) 2025-04-12 22:50:37 -07:00
Yineng Zhang
7d3b7c87f5 fix: determine if flashinfer is installed (#5336) 2025-04-12 19:59:13 -07:00
tianlian yi
bc92107b03 Support server based rollout in Verlengine (#4848) 
Co-authored-by: Jin Pan <jpan236@wisc.edu>
Co-authored-by: Chayenne <zhaochen20@outlook.com>
Co-authored-by: Jinn <47354855+jhinpan@users.noreply.github.com>
 2025-04-12 10:07:52 -07:00
Xiaoyu Zhang
3e4794aad8 refine fused_moe tuning docs (#5294) 2025-04-12 10:01:13 -07:00
Xiaoyu Zhang
690ec20587 Delete python/sglang/srt/layers/moe/fused_moe_triton/configs/E=257,N=… (#5321) 2025-04-12 10:00:03 -07:00
Yineng Zhang
57de7c6b5f feat: use fa3 mla by default on hopper (#5210) 
Co-authored-by: yundai424 <yundai424@gmail.com>
Co-authored-by: hebiao064 <hebiaobuaa@gmail.com>
 2025-04-12 01:09:25 -07:00
Qingquan Song
aea98512a8 Fix fa3 window size setup (#5316) 2025-04-11 23:37:52 -07:00
lambert0312
1b1b47a949 Fix w8a8_int8 model shared experts fusion load weights error (#5120) 2025-04-11 23:33:51 -07:00
Mick
34ef6c8135 [VLM] Adopt fast image processor by default (#5065) 2025-04-11 21:46:58 -07:00
Yineng Zhang
611720919d fix: use deepgemm only on hopper (#5310) 2025-04-11 20:48:24 -07:00
Yineng Zhang
f774a0d275 feat: add blackwell Dockerfile (#5302) 2025-04-11 13:08:53 -07:00
Xiaoyu Zhang
60bcbf2a35 remove moe_align_block_size torch.zeros in small batch/expert mode (#5298) 2025-04-11 12:13:55 -07:00
Yusong Gao
c35dcfdb30 [PD] fix: skip warmup request in disaggregation mode to prevent crash on timeout (#5292) 2025-04-11 23:03:07 +08:00
Mick
e53a0b3d5b [fix] fix mrope positions not picked up (#5265) 2025-04-11 01:29:45 -07:00
Cheng Wan
038bc5d521 Support --enable-llama4-multimodal (#5254) 2025-04-11 01:24:14 -07:00
Chang Su
aee62d744b Optimize GPU memory usage in FlashAttentionBackend's strided indexing (#5262) 
Co-authored-by: ch-wan <cwan39@gatech.edu>
 2025-04-11 00:34:17 -07:00
fzyzcjy
cd7e32e2cb Optimize attention in llama4 (#5127) 2025-04-11 00:32:41 -07:00
HAI
8879944800 ROCm/AITER CK_MoE: update 2-stage kernels & support both Activations (#5228) 2025-04-10 18:19:57 -07:00
Richard Zou
a879811c4b Fix torch.compile cacheing (#5259) 
Co-authored-by: zhyncs <me@zhyncs.com>
 2025-04-10 18:08:45 -07:00
Ke Bao
1078396f47 Update deps for mllama4 (#5215) 2025-04-10 09:12:44 -07:00
Teng Ma
7e4f72dd8c [PD] Add get_contiguous_buf_infos interface for MLATokenToKVPool (#5204) 2025-04-10 20:05:34 +08:00
Teng Ma
4c31ae9f6d [PD] Support KV transfer with mooncake (#4880) 
Signed-off-by: Shangming Cai <caishangming@linux.alibaba.com>
Co-authored-by: Shangming Cai <caishangming@linux.alibaba.com>
Co-authored-by: Xuchun Shang <xuchun.shang@linux.alibaba.com>
Co-authored-by: shangmingc <csmthu@gmail.com>
 2025-04-10 14:23:23 +08:00
Xiaoyu Zhang
f730362ee2 reduce moe_align_block_size_kernel small batch mode overhead (#5086) 2025-04-09 17:59:35 -07:00
fzyzcjy
e3c4bd3153 Fix DeepSeek error when using DeepEP mode (#5190) 2025-04-09 17:43:22 -07:00