Commits · 4060ed37cb67262b0cc7af2bcbbdf37ba12d3501 · change / sglang · GitLab

24 Oct, 2025 20 commits
- Refactoring GLM-4.5 and GLM-4.5V related implementations (#11800) · 4060ed37
  Yuxuan Zhang authored Oct 24, 2025
  
  4060ed37
- Tiny cleanup send_single (#12056) · 2342605e
  fzyzcjy authored Oct 24, 2025
  
  2342605e
- [router] Add mTLS Support for Router-to-Worker Communication (#12019) · dbf17a83
  Simo Lin authored Oct 23, 2025
```
Co-authored-by: Chang Su <chang.s.su@oracle.com>
```
  dbf17a83
- Install numactl in Dockerfile for GH200/GB200/GB300 (#11853) · 0f0c430e
  fzyzcjy authored Oct 24, 2025
  
  0f0c430e
- fix: the hardcode hf repo name comparison for deepseek-ocr (#12031) · 8e797a47
  Rain Jiang authored Oct 23, 2025
```
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
```
  8e797a47
- Add gguf dependency for cpu/xpu (#12041) · aa3003f1
  Zaili Wang authored Oct 24, 2025
  
  aa3003f1
- Opt MHA chunked prefix: merge prefix and extend kv cache to run mha once (#10953) · 4793ec7d
  Yongfei Xu authored Oct 24, 2025
  
  4793ec7d
- fix: fix MMMU loading issue (#11759) · 92009bd2
  Zaili Wang authored Oct 24, 2025
```
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
```
  92009bd2
- Revert "[Fix] Fix lint to pass CI" (#12042) · 4ef981e2
  Baizhou Zhang authored Oct 23, 2025
  
  4ef981e2
- [Fix] Fix lint to pass CI (#12037) · 69ed8b67
  Baizhou Zhang authored Oct 23, 2025
  
  69ed8b67
- support more model in piecewise cuda graph (#11745) · 1801cd19
  narutolhy authored Oct 23, 2025
  
  1801cd19
- Revert "lang: support direct video inference" (#12038) · ffc722a6
  Lianmin Zheng authored Oct 23, 2025
  
  ffc722a6
- Fix(security): block unsafe pickle deserialization to mitigate CVE-2025-10164 (#11909) · 49afb3d9
  thelongestusernameofall authored Oct 24, 2025
```
Co-authored-by: Chengxing Xie <xiechengxing34@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
```
  49afb3d9
- Use flashinfer_trtllm moe runner backend to gain around 10% perf on b200 fp8 dpsk (#11816) · f80371ff
  b8zhong authored Oct 23, 2025
  
  f80371ff
- Refactor Triton-kernel MoE runner integration (#11795) · 62eff37b
  Jonah Bernard authored Oct 23, 2025
  
  62eff37b
- Enable Llama 4 + TRTLLM MHA (#12003) · 47e12e08
  b8zhong authored Oct 23, 2025
  
  47e12e08
- lang: support direct video inference (#9936) · 823b4429
  Mick authored Oct 24, 2025
```
Co-authored-by: Lianmin Zheng <lianminzheng@gmail.com>
```
  823b4429
- [8/n] decouple quantization impl from vllm dependency - gguf srt (#11964) · 14a4d80e
  Fan Yin authored Oct 24, 2025
```
Co-authored-by: Peng Zhang <zhuangsen.zp@antgroup.com>
```
  14a4d80e
- chore: bump SGLang version to 0.5.4 (#12027) · 1053e1be
  sglang-bot authored Oct 24, 2025
```
Co-authored-by: sglang-bot <sglang-bot@users.noreply.github.com>
```
  1053e1be
- Fixed aarch64 flash-mla (#12009) · 9a71500c
  nvjullin authored Oct 24, 2025
  
  9a71500c
23 Oct, 2025 20 commits
- [router] Add builder pattern for RouterConfig with zero duplication (#12030) · 6d6e24bc
  Simo Lin authored Oct 23, 2025
  
  6d6e24bc
- Update Github action title for kernel build (#12029) · 2c057fbf
  Kangyan-Zhou authored Oct 23, 2025
  
  2c057fbf
- Fix mamba radix cache eviction logic in `alloc_req_slots` (#11616) · dbd9435d
  Roger Young authored Oct 24, 2025
```
Signed-off-by: rogeryoungh <rogeryoungh@foxmail.com>
```
  dbd9435d
- Revert "[ROCm] Remove vLLM rope dependency & use AITER impl" (#12028) · 8ae9d4bb
  b8zhong authored Oct 23, 2025
  
  8ae9d4bb
- Log iteration # for prefill and decode (#9366) · 1c304aa9
  Nicolas Castet authored Oct 23, 2025
  
  1c304aa9
- model: support deepseek-ocr (#11891) · 770529a7
  Mick authored Oct 24, 2025
```
Co-authored-by: yhyang201 <47235274+yhyang201@users.noreply.github.com>
Co-authored-by: yhyang201 <yhyang201@gmail.com>
Co-authored-by: Shi Shuai <126407087+shuaills@users.noreply.github.com>
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
```
  770529a7
- Add AWQ quantization support for NPU. (#10158) · 39c237f0
  ErvinXie authored Oct 24, 2025
```
Co-authored-by: Alisehen <814073252@qq.com>
Co-authored-by: Yaochen Han <48639761+Alisehen@users.noreply.github.com>
Co-authored-by: Zhengda Qin <zhengdqin@gmail.com>
```
  39c237f0
- [router][CI] Clean up imports and prints statements in sgl-router/py_test (#12024) · 28b8a406
  Chang Su authored Oct 23, 2025
  
  28b8a406
- ci: fix night-ci with push retry mechanism (#11765) · 8bd26dd4
  Mick authored Oct 24, 2025
  
  8bd26dd4
- [Auto Sync] Update test_deterministic_utils.py (20251023) (#12022) · ab07cd3e
  Lianmin Zheng authored Oct 23, 2025
```
Co-authored-by: Stefan He <hebiaobuaa@gmail.com>
```
  ab07cd3e
- Feature/nano v2 offline modelopt fp8 and nvfp4 (#12018) · a9849683
  Netanel Haber authored Oct 23, 2025
```
Signed-off-by: Netanel Haber <58652339+netanel-haber@users.noreply.github.com>
```
  a9849683
- [router] change ci names and update log level in ci (#12021) · a4b637d8
  Simo Lin authored Oct 23, 2025
  
  a4b637d8
- [Feature] Support loading weights from ckpt engine worker (#11755) · 96a5e4dd
  Teng Ma authored Oct 24, 2025
```
Signed-off-by: Yang Kaiyong <yangkaiyong.yky@antgroup.com>
Signed-off-by: Cruz Zhao <CruzZhao@linux.alibaba.com>
Signed-off-by: Xuchun Shang <xuchun.shang@gmail.com>
Co-authored-by: Yang Kaiyong <yangkaiyong.yky@antgroup.com>
Co-authored-by: Cruz Zhao <CruzZhao@linux.alibaba.com>
Co-authored-by: Xuchun Shang <xuchun.shang@gmail.com>
Co-authored-by: Shangming Cai <csmthu@gmail.com>
```
  96a5e4dd
- [Fix] memory leak by overlap + retract (#11981) · b0b4f716
  cctry authored Oct 23, 2025
```
Co-authored-by: Liangsheng Yin <lsyincs@gmail.com>
```
  b0b4f716
- Revert "Support nvidia/NVIDIA-Nemotron-Nano-9B-v2-FP8/NVFP4" (#12015) · 6c18addb
  Liangsheng Yin authored Oct 23, 2025
  
  6c18addb
- Move memory runtime checker to mixin class (#12014) · 32852fe9
  Liangsheng Yin authored Oct 23, 2025
  
  32852fe9
- [Router] Consolidate ConnectionMode enum to core module (#11937) · 53c2934d
  Arthur Cheng authored Oct 23, 2025
  
  53c2934d
- [router] Add comprehensive E2E tests for Response API (#11988) · e321c971
  Keyang Ru authored Oct 23, 2025
  
  e321c971
- Support nvidia/NVIDIA-Nemotron-Nano-9B-v2-FP8/NVFP4 (#11866) · d6fee73d
  Netanel Haber authored Oct 23, 2025
  
  d6fee73d
- Support overlap-spec-v2 with trtllm_mla attention backend (#11821) · 36a4cad7
  Qiaolin Yu authored Oct 23, 2025
  
  36a4cad7