Commits · b6c14ec0b4f3d7f744c734a3835298b3242a2b90 · change / sglang · GitLab

26 Aug, 2025 21 commits
- add `response_format` support for `completion` API (#9665) · b6c14ec0
  cicirori authored Aug 27, 2025
  
  b6c14ec0
- HiCache Storage fix host memory leak (#9648) · 43de1d73
  Zhiqiang Xie authored Aug 26, 2025
  
  43de1d73
- BugFix(hicache): Fix host indices out of bound error (#9637) · 79ce3688
  hzh0425 authored Aug 27, 2025
```
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
```
  79ce3688
- Install py-spy by default for containers for easier debugging (#9649) · 44ffe2cb
  fzyzcjy authored Aug 27, 2025
  
  44ffe2cb
- [doc] add kimik2 --tool-call-parser (#9647) · 1a0896e9
  Xiaotong Jiang authored Aug 26, 2025
  
  1a0896e9
- [router] add token bucket rate limiter (#9656) · 90313fb0
  Chang Su authored Aug 26, 2025
  
  90313fb0
- [router] address worker load tracking consistency (#9523) · 3578eb1e
  Simo Lin authored Aug 26, 2025
```
Co-authored-by: fzyzcjy <5236035+fzyzcjy@users.noreply.github.com>
```
  3578eb1e
- Fix kimi k2 function calling format (#9606) · 0936c766
  Xiaotong Jiang authored Aug 26, 2025
  
  0936c766
- fix: allow user to specify function as role (#9635) · 0ef583b7
  GavinZhu-GMI authored Aug 26, 2025
  
  0ef583b7
- Add reasoning_effort param in TiktokenTokenizer.apply_chat_template (#9630) · f7881a27
  Liu Shaohui authored Aug 26, 2025
```
Co-authored-by: Shaohui Liu <liushaohui3@xiaomi.com>
Co-authored-by: Xinyuan Tong <115166877+JustinTong0323@users.noreply.github.com>
```
  f7881a27
- [docs] Update README with additional highlights and resources for SGLang x AMD SF Meetup (#9640) · fdff3167
  Mingyi authored Aug 26, 2025
  
  fdff3167
- Fix lint for router (#9636) · cbc0e4d7
  Stefan He authored Aug 26, 2025
  
  cbc0e4d7
- model: Support nvidia/Llama-3_1-Nemotron-Ultra-253B-v1 (#9301) · 4cd08dc5
  Netanel Haber authored Aug 26, 2025
  
  4cd08dc5
- [new feat] ascend backend support fia fusion kernel (#8328) · f92b729d
  ZhengdQin authored Aug 26, 2025
```
Co-authored-by: Even Zhou <even.y.zhou@outlook.com>
```
  f92b729d
- [router] add ut for mistral, llama, pythonic, and streaming tool parser (#9632) · e2e378ca
  Simo Lin authored Aug 25, 2025
```
Co-authored-by: Chang Su <chang.s.su@oracle.com>
```
  e2e378ca
- [router] add llama tool parser (#9629) · dc1decc6
  Simo Lin authored Aug 25, 2025
```
Co-authored-by: Chang Su <chang.s.su@oracle.com>
```
  dc1decc6
- [router] add pythonic parser (#9628) · 03680f33
  Simo Lin authored Aug 25, 2025
```
Co-authored-by: Chang Su <chang.s.su@oracle.com>
```
  03680f33
- [router] add qwen tool parser (#9623) · d4c5e534
  Simo Lin authored Aug 25, 2025
```
Co-authored-by: Chang Su <chang.s.su@oracle.com>
```
  d4c5e534
- [router] add mistral tool parser (#9622) · 817c62a0
  Simo Lin authored Aug 25, 2025
```
Co-authored-by: Chang Su <chang.s.su@oracle.com>
```
  817c62a0
- Improve bench_one_batch_server script (#9608) · 0ff72419
  Liangsheng Yin authored Aug 26, 2025
```
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
```
  0ff72419
- [Fix] HiCache Bugfix & Mooncake Error Handling Enhance (#8901) · 80dc76e1
  ykwd authored Aug 26, 2025
```
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
```
  80dc76e1
25 Aug, 2025 19 commits
- [docs] Refactor, remove compiled results and add gpt-oss (#9613) · 9b08d975
  Chayenne authored Aug 25, 2025
```
Co-authored-by: zhaochenyang20 <zhaochenyang20@gmail.com>
```
  9b08d975
- Fix Harmony reasoning parser for and auto-separation for gpt-oss models (#9190) · a0a77d93
  Jonas authored Aug 26, 2025
```
Co-authored-by: Chang Su <chang.s.su@oracle.com>
Co-authored-by: Chayenne <zhaochen20@outlook.com>
Co-authored-by: zhaochenyang20 <zhaochenyang20@gmail.com>
Co-authored-by: minleminzui <2969413251@qq.com>
Co-authored-by: maocheng23 <maocheng@berkeley.edu>
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
```
  a0a77d93
- Fix GLM45v launch server cuda torch compile bug (#9554) · 24a8cee6
  Binyao Jiang authored Aug 25, 2025
  
  24a8cee6
- Fix GLM45 tool call multi-turn bug (#9500) · 3affa9dc
  Binyao Jiang authored Aug 25, 2025
  
  3affa9dc
- [Performance] Batch Send from Tokenizer Manager. (#9436) · ea0696b9
  Sundara Raman Ramachandran authored Aug 25, 2025
  
  ea0696b9
- [Doc] add LWS(LeaderWorkerSet) use case in sgl-router README (#9568) · 3aec3d4f
  Bruce-x-1997 authored Aug 25, 2025
```
Co-authored-by: bruce.xu <bruce.xu@gmicloud.ai>
```
  3aec3d4f
- chore: bump v0.5.1.post2 (#9592) · e3e97a12
  Yineng Zhang authored Aug 25, 2025
  
  e3e97a12
- chore: update config (#9591) · 05106867
  Yineng Zhang authored Aug 25, 2025
  
  05106867
- Tiny fix wrong comments (#9589) · 9dcdf5da
  fzyzcjy authored Aug 25, 2025
  
  9dcdf5da
- fix: resolve tuning fused moe issue (#9587) · f8b757bc
  Yineng Zhang authored Aug 25, 2025
  
  f8b757bc
- fix: revert #8593 (#9581) · ebd9dbe7
  Yineng Zhang authored Aug 25, 2025
  
  ebd9dbe7
- chore: upgrade flashinfer 0.2.14.post1 (#9578) · 938e986e
  Yineng Zhang authored Aug 25, 2025
  
  938e986e
- bugfix for undefined logging functions in HarmonyBrowserTool & HarmonyPythonTool (#9229) · 17d5eda8
  Yuhao Zhou authored Aug 25, 2025
  
  17d5eda8
- Offload tensors by sharding on GPU (#9536) · 71a7f1d8
  fzyzcjy authored Aug 25, 2025
  
  71a7f1d8
- Reintroduce memory usage fix (#9535) · 433266c1
  fzyzcjy authored Aug 25, 2025
  
  433266c1
- Update CUTLASS 4.2 & Enable K-Major Scale Factor for SM90 FP8 Blockwise Group GEMM (#9559) · fda47926
  Qi Yuhang authored Aug 25, 2025
  
  fda47926
- remove redundant rank0_log function. (#9560) · a0b22f2f
  miter authored Aug 25, 2025
```
Co-authored-by: linhuang <linhuang@ruijie.com.cn>
```
  a0b22f2f
- [PD] Improve disaggregation metrics output: update the metrics to keep... · b5c6529e
  SCDESPERTATE authored Aug 25, 2025
```
[PD] Improve disaggregation metrics output: update the metrics to keep reflecting real stats (#7317)
```
  b5c6529e
- fix: Update OpenAI client base URL in documentation (#9576) · ca4b86c5
  Xinyuan Tong authored Aug 25, 2025
  
  ca4b86c5