Commits · 1e0806f30b99205bffe453b0caef77bc21e99baa · change / sglang · GitLab

19 Apr, 2025 2 commits
- Fix DeepGEMM masked cannot be run on groups not being multiple or 4 (#5340) · 1e0806f3
  fzyzcjy authored Apr 19, 2025
  
  1e0806f3
- chore: upgrade sgl-kernel 0.0.9.post2 (#5540) · 2c11f9c2
  Yineng Zhang authored Apr 18, 2025
  
  2c11f9c2
18 Apr, 2025 14 commits
- Revert "Avoid computing lse in Ragged Prefill when there's no prefix.… (#5544) · a6f892e5
  Yineng Zhang authored Apr 18, 2025
  
  a6f892e5
- fix util import (#5542) · 08b518d5
  Yineng Zhang authored Apr 18, 2025
  
  08b518d5
- [Model] Adding Qwen3 and Qwen3MoE (#4693) · 4db463b1
  yhyang201 authored Apr 19, 2025
  
  4db463b1
- Avoid computing lse in Ragged Prefill when there's no prefix. (#5476) · bfa39224
  Wenxuan Tan authored Apr 18, 2025
```
Co-authored-by: Baizhou Zhang <sobereddiezhang@gmail.com>
```
  bfa39224
- fix bug of VLLM_AVAILABLE not defined (#5497) · e465b08d
  liwenju0 authored Apr 18, 2025
  
  e465b08d
- fix kimi vl running bug after rebase main (#5461) · bed05878
  Xiaoyu Zhang authored Apr 18, 2025
  
  bed05878
- use sglang_per_token_group_quant_fp8 from sgl-kernel instead of trion kernel (#5473) · b2a189dd
  strgrb authored Apr 18, 2025
```
Co-authored-by: Zhang Kaihong <zhangkaihong.zkh@alibaba-inc.com>
```
  b2a189dd
- chore: bump sgl-kernel 0.0.9.post2 (#5518) · f28d8299
  Yineng Zhang authored Apr 17, 2025
  
  f28d8299
- Sgl kernel fused_moe_gate support n_shared_experts (#5440) · 8e09b370
  Xiaoyu Zhang authored Apr 18, 2025
  
  8e09b370
- Introduce moe_dense_tp_size to fix dense layer errors in DeepSeek V3 + 4x8xH100 (#4836) · 53dcf388
  fzyzcjy authored Apr 18, 2025
  
  53dcf388
- Configuration qwen2_moe.py - qkv_bias now in transformers (#5512) · 1effba4c
  Michael Feil authored Apr 17, 2025
  
  1effba4c
- [docs] Fix several consistency issues in sampling_params.md (#5373) · a0fc5bc1
  Michael Yao authored Apr 18, 2025
```
Signed-off-by: windsonsea <haifeng.yao@daocloud.io>
Co-authored-by: Baizhou Zhang <sobereddiezhang@gmail.com>
```
  a0fc5bc1
- Fix: fix the exception 'the memory capacity is unbalanced. Some GPUs … (#5426) · 27e9538a
  mlmz authored Apr 18, 2025
```
Co-authored-by: ocss884 <ocss.lin@gmail.com>
```
  27e9538a
- Fix: Incorrect parameters passed to forward_batch_generation (#5506) (#5511) · 211c7b31
  u4lr451 authored Apr 18, 2025
  
  211c7b31
17 Apr, 2025 13 commits
- [Feat] Update sgl-kernel flashinfer to latest main version (#5500) · c08a717c
  PGFLMG authored Apr 18, 2025
```
Co-authored-by: zhyncs <me@zhyncs.com>
```
  c08a717c
- Doc: fix problems of the 'Execute Notebooks / run-all-notebooks' ci caused by... · f13d65a7
  mlmz authored Apr 18, 2025
```
Doc: fix problems of the 'Execute Notebooks / run-all-notebooks' ci caused by the unstability of deepseek-ai/DeepSeek-R1-Distill-Qwen-7B (#5503)
```
  f13d65a7
- [Bug fix] use correct func path in deepseek (#5496) · 06d0a3d9
  Xuchun Shang authored Apr 17, 2025
```
Signed-off-by: Xuchun Shang <xuchun.shang@linux.alibaba.com>
```
  06d0a3d9
- Fix a link in sgl-kernel/README.md (#5493) · 22c2a79d
  Michael Yao authored Apr 17, 2025
  
  22c2a79d
- Refactor DeepSeek decoder layer branches (#5205) · 8beb356f
  fzyzcjy authored Apr 17, 2025
  
  8beb356f
- Enable local attention during decode (#5479) · c776234b
  Chang Su authored Apr 17, 2025
  
  c776234b
- Feat/support encoder model (like bert) (#4887) · 3bface15
  woodx authored Apr 17, 2025
  
  3bface15
- Deprecate enable-flashinfer-mla and enable-flashmla (#5480) · 6fb29ffd
  Baizhou Zhang authored Apr 17, 2025
  
  6fb29ffd
- Deprecate disable-mla (#5481) · 4fb05583
  Baizhou Zhang authored Apr 17, 2025
  
  4fb05583
- Add test for flash_attn_varlen_func kernel (#5484) · 81c89111
  Baizhou Zhang authored Apr 17, 2025
  
  81c89111
- Update attention_backend.md: plural form (#5489) · 92d1561b
  Didier Durand authored Apr 17, 2025
  
  92d1561b
- [Model Support] unsloth/Phi-4-mini bnb model (#4982) · 8f783c19
  eigen authored Apr 16, 2025
```
Co-authored-by: yhyang201 <yhyang201@gmail.com>
Co-authored-by: Liangsheng Yin <hnyls2002@gmail.com>
Co-authored-by: Chayenne <zhaochen20@outlook.com>
Co-authored-by: Yineng Zhang <me@zhyncs.com>
```
  8f783c19
- [verl] Modify the update_weights func to align with verl's resharding (#5345) · 90faf901
  BearBiscuit authored Apr 17, 2025
```
Co-authored-by: Chayenne <zhaochen20@outlook.com>
```
  90faf901
16 Apr, 2025 11 commits
- Clean up imports (#5467) · 177320a5
  Lianmin Zheng authored Apr 16, 2025
  
  177320a5
- add multi-lora feature in README.md (#5463) · d7bc19a4
  Ying Sheng authored Apr 16, 2025
  
  d7bc19a4
- Update cutlass dependency. (#5447) · 85ec0440
  Elfie Guo authored Apr 15, 2025
  
  85ec0440
- [doc] Update benchmark_and_profiling.md (#5449) · 06a1656e
  Xiaoyu Zhang authored Apr 16, 2025
  
  06a1656e
- Fix several minor issues in PD disaggregation (#5444) · 6aca5834
  Cheng Wan authored Apr 15, 2025
  
  6aca5834
- chore: bump v0.4.5.post1 (#5445) · 5b5c7237
  Yineng Zhang authored Apr 15, 2025
  
  5b5c7237
- Support MHA with chunked prefix cache for DeepSeek chunked prefill (#5113) · a42736bb
  Baizhou Zhang authored Apr 15, 2025
  
  a42736bb
- [Bug fix] need record start time in pd mode (#5425) · dd83e7e9
  ybyang authored Apr 16, 2025
  
  dd83e7e9
- [Minor] Move torch.compile patch to a better place (#5397) · 0769b14b
  Lianmin Zheng authored Apr 15, 2025
  
  0769b14b
- [Docs] Update start/install.md (#5398) · b64b88e7
  Michael Yao authored Apr 16, 2025
  
  b64b88e7
- Support BNB quantization for llama/mllama (#5038) · bc24205b
  ryang authored Apr 16, 2025
```
Co-authored-by: Yuhao Yang <yyh073@foxmail.com>
```
  bc24205b