Commits · 61a4680494902efeef00ce7559dca3869bda3881 · change / sglang

"docs/git@developer.sourcefind.cn:change/sglang.git" did not exist on "ac5b78baf665a6c479bcc1c3a29fc3874cbf718d"

08 Aug, 2025 14 commits
- chore: bump v0.5.0rc0 (#8959) · 9020f7fc
  Yineng Zhang authored Aug 08, 2025
  
  9020f7fc
- [RL] fix skip_server_warmup and rl health_generate logic (#8757) · dd650e0e
  Zilin Zhu authored Aug 08, 2025
  
  dd650e0e
- Revert "Support Multi Process Tokenizer Manager" (#8960) · a9471542
  Lianmin Zheng authored Aug 08, 2025
  
  a9471542
- Simple prefetch policy (#8692) · e2fd2b9c
  pansicheng authored Aug 08, 2025
  
  e2fd2b9c
- Support Multi Process Tokenizer Manager (#6555) · 7490e3f6
  ybyang authored Aug 08, 2025
```
Signed-off-by: ybyang <ybyang7@iflytek.com>
Signed-off-by: huanglong <huanglong@linux.alibaba.com>
Co-authored-by: lw9527 <952799980@qq.com>
Co-authored-by: huanglong <huanglong@linux.alibaba.com>
Co-authored-by: Huang Long <121648372+LLLL114@users.noreply.github.com>
```
  7490e3f6
- add zai-org/GLM-4.5-Air-FP8 model into nightly CI (#8894) · 6ee6619b
  Minglei Zhu authored Aug 08, 2025
  
  6ee6619b
- [NVIDIA] Fix missing `get_col_major_tma_aligned_tensor` for Blackwell deepgemm in EpMoE (#8955) · b4c9f38a
  Kaixi Hou authored Aug 08, 2025
  
  b4c9f38a
- Add ernie4.py for ERNIE-4.5 (#7657) · 11325474
  Wenbo Yang authored Aug 08, 2025
  
  11325474
- Expert Parallelism for GPT-OSS (#8944) · 1d24db83
  Cheng Wan authored Aug 08, 2025
  
  1d24db83
- minor: global workspace buffer for trtllm-gen mha from flashinfer (#8952) · 08fab2b0
  eigen authored Aug 08, 2025
  
  08fab2b0
- Better optimization log for gpt-oss model (#8953) · 0d1e27a0
  Xiaoyu Zhang authored Aug 08, 2025
  
  0d1e27a0
- Reduce scheduler recv requests overhead (#8947) · 774b47f3
  fzyzcjy authored Aug 08, 2025
  
  774b47f3
- Fix enable flashinfer mxfp4 moe bf16 check (#8950) · 76915d68
  Xiaoyu Zhang authored Aug 08, 2025
  
  76915d68
- Enhancements for bench_one_batch (#8703) · ed0a3dd5
  Zaili Wang authored Aug 08, 2025
```
Co-authored-by: root <root@gnr630186.jf.intel.com>
```
  ed0a3dd5
07 Aug, 2025 14 commits
- correct the tp_plan logic (#8850) · d3be9710
  Stefan He authored Aug 07, 2025
  
  d3be9710
- fix: reasoning parser when request have enable_thinking flag (#8933) · 3e7ff1ab
  Xinyuan Tong authored Aug 07, 2025
```
Signed-off-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
```
  3e7ff1ab
- remove vllm fp8quant from fp8.py (#8937) · aaf0ad8c
  Stefan He authored Aug 07, 2025
  
  aaf0ad8c
- fix: use openai 1.99.1 (#8927) · 4bf6e5a6
  Yineng Zhang authored Aug 07, 2025
  
  4bf6e5a6
- Fix hopper launch gpt-oss model illegal memory (#8908) · 3ae33fcd
  Xiaoyu Zhang authored Aug 08, 2025
  
  3ae33fcd
- [Perf] Auto enable best flashinfer mxfp4 kernel in b200 (#8898) · 47824c14
  Xiaoyu Zhang authored Aug 07, 2025
  
  47824c14
- Disable gemma3 for SWA due to CUDA illegal memory access error (#8895) · c36a6693
  Xinyuan Tong authored Aug 07, 2025
```
Signed-off-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
```
  c36a6693
- [CPU] Fix fallback allgather issue (#8041) · 62f8eb48
  blzheng authored Aug 07, 2025
  
  62f8eb48
- [Feat] QWen-1M context support[2/2]: Update block sparse attention backend (#5949) · b7cd7430
  PGFLMG authored Aug 07, 2025
  
  b7cd7430
- [Feature][Multimodal] Implement LRU cache for multimodal embeddings (#8292) · 2d120f8b
  Zheng Wengang authored Aug 07, 2025
```
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>
```
  2d120f8b
- Enables force reasoning based on chat template for Qwen3-Thinking (#8369) · 3fa3c6cd
  Xinyuan Tong authored Aug 06, 2025
```
Signed-off-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
Signed-off-by: Xinyuan Tong <justinning0323@outlook.com>
Co-authored-by: Chang Su <csu272@usc.edu>
```
  3fa3c6cd
- Support GPU pinning for LoRA (#8697) · 6210e2c4
  Lifu Huang authored Aug 06, 2025
  
  6210e2c4
- feat: openai oss attention sink support with trtllm-gen backend #8825 (#8834) · 6ad6c8c9
  eigen authored Aug 06, 2025
```
Co-authored-by: averyhuang <averyh@nvidia.com>
```
  6ad6c8c9
- fix glm4 moe (#8883) · 5b6acc14
  Cheng Wan authored Aug 06, 2025
  
  5b6acc14
06 Aug, 2025 12 commits
- add flashinfer mxfp4 (#8847) · 4373df55
  Xiaoyu Zhang authored Aug 07, 2025
  
  4373df55
- Use reduce scatter for DP (#8539) · c0e84297
  Trevor Morris authored Aug 06, 2025
  
  c0e84297
- Support v1/responses and use harmony in serving_chat (#8837) · 92cc32d9
  Chang Su authored Aug 06, 2025
```
Signed-off-by: Xinyuan Tong <justinning0323@outlook.com>
Signed-off-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
Co-authored-by: Xinyuan Tong <justinning0323@outlook.com>
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
```
  92cc32d9
- [NVIDIA] Fix num_experts in modelopt_quant (#8811) · 288ae41f
  Shu Wang authored Aug 06, 2025
  
  288ae41f
- Optimize triton swa kernel by skipping computation (#8860) · 0475448e
  Ke Bao authored Aug 06, 2025
  
  0475448e
- Refine naming (#8868) · 399e7ec8
  Ke Bao authored Aug 06, 2025
  
  399e7ec8
- Support mxfp4 for GPT-OSS (#8843) · 168033d5
  Ying Sheng authored Aug 06, 2025
```
Co-authored-by: Co-author fzyzcjy <ch271828n@outlook.com>
Co-authored-by: fzyzcjy <5236035+fzyzcjy@users.noreply.github.com>
Co-authored-by: zhuofan1123 <zhuofanl@nvidia.com>
Co-authored-by: liz-badada <jinyanc@nvidia.com>
Co-authored-by: xutizhou <xutingz@nvidia.com>
Co-authored-by: linhu-nv <linhu@nvidia.com>
```
  168033d5
- [2/3] Optimize Slime Update Weights: Avoid GPU-to-CPU Device Sync when update... · cbbb7383
  Stefan He authored Aug 05, 2025
```
[2/3]  Optimize Slime Update Weights: Avoid GPU-to-CPU Device Sync when update expert weights (#8753)
```
  cbbb7383
- [1/3] Optimize Slime Update Weights: Remove QWen3MOE Load Weight Overhead (#8751) · 89588179
  Stefan He authored Aug 05, 2025
  
  89588179
- [Feature] improve TBO: two chunk overlap (#8144) · ca47e24f
  HouseWest authored Aug 06, 2025
  
  ca47e24f
- Support bailing moe (#8680) · d26ca84f
  Praneth Paruchuri authored Aug 06, 2025
  
  d26ca84f
- Turn off hybrid cache by default (#8839) · 8128e08d
  Ke Bao authored Aug 06, 2025
  
  8128e08d