Commits · a0557642eab9e1375df20e05d132890b95163d0a · change / sglang · GitLab

09 Oct, 2025 5 commits
- [router][lint] Add unused_qualifications to cargo lint warnings (#11366) · a0557642
  Chang Su authored Oct 08, 2025
  
  a0557642
- [router] Refactor OpenAI router: split monolithic file and move location (#11359) · 84768d10
  Keyang Ru authored Oct 08, 2025
  
  84768d10
- [router][grpc] disable health check generation and increase timeout (#11353) · 368fd206
  Simo Lin authored Oct 08, 2025
  
  368fd206
- [Generative Score API] Multi-Item scoring with custom attention mask. (#10979) · 53bd00d9
  Sundara Raman Ramachandran authored Oct 08, 2025
  
  53bd00d9
- [Auto Sync] Update scheduler.py (20251009) (#11350) · e22b13c5
  Yineng Zhang authored Oct 08, 2025
```
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Junxiong Wang <junxiong@together.ai>
```
  e22b13c5
08 Oct, 2025 25 commits
- fix: fix revision for sgl-flash-attn in sgl-kernel (#11327) · a3c2ea44
  Mick authored Oct 09, 2025
  
  a3c2ea44
- [router][grpc] Add dependencies in Cargo.toml to support chat template rendering (#11342) · fccac7d1
  Chang Su authored Oct 08, 2025
  
  fccac7d1
- [router] Support history management using conversation (#11339) · 7ac6b900
  Keyang Ru authored Oct 08, 2025
  
  7ac6b900
- [router] Fix all unused_qualifications (#11341) · a1080b72
  Chang Su authored Oct 08, 2025
  
  a1080b72
- [router][grpc] Cleanup debug logs in grpc_server and grpc_router (#11340) · a65ca739
  Chang Su authored Oct 08, 2025
  
  a65ca739
- [router] improve reasoning parser lock and reduce req cloning (#11336) · 677aa0e2
  Simo Lin authored Oct 08, 2025
  
  677aa0e2
- [router] refactor generate to use new pipeline arch (#11323) · 01c9ee1a
  Simo Lin authored Oct 08, 2025
  
  01c9ee1a
- model: Support Hybrid Mamba2 NemotronHForCausalLM (nvidia/NVIDIA-Nemotron-Nano-9B-v2) (#10909) · d6837aea
  Netanel Haber authored Oct 08, 2025
```
Signed-off-by: Netanel Haber <nhaber@nvidia.com>
```
  d6837aea
- [CI] improve disaggregation CI. (#11264) · c882b5ae
  Liangsheng Yin authored Oct 08, 2025
```
Signed-off-by: Shangming Cai <csmthu@gmail.com>
Co-authored-by: Shangming Cai <csmthu@gmail.com>
```
  c882b5ae
- benchmark: enhance configurable multimodal benchmarking in bench_serving (#9812) · e3bb7f5a
  Kevin Xiang Li authored Oct 08, 2025
```
Co-authored-by: Xiang (Kevin) Li <lik@nvidia.com>
Co-authored-by: Xinyuan Tong <115166877+JustinTong0323@users.noreply.github.com>
```
  e3bb7f5a
- Support LoRA in bench_serving oai interface (#11318) · 92473e2e
  Lifu Huang authored Oct 08, 2025
  
  92473e2e
- fix(decode): adjust ServerArgs import to explicit module path (#11007) · 6c0bb327
  JinYan Su authored Oct 08, 2025
  
  6c0bb327
- [Doc] Update mooncake nvlink transport doc for PD disaggregation (#11321) · 0a7c4bde
  Shangming Cai authored Oct 08, 2025
```
Signed-off-by: Shangming Cai <csmthu@gmail.com>
```
  0a7c4bde
- [2/2] Support MHA prefill with FlashAttention 4. (#10937) · edefab0c
  Lifu Huang authored Oct 08, 2025
```
Co-authored-by: Hieu Pham <hyhieu@gmail.com>
```
  edefab0c
- Skip weight loading in deepgemm compilation (#11312) · 97cd38e5
  Cheng Wan authored Oct 07, 2025
  
  97cd38e5
- [8/N] MoE Refactor: deprecate `EPMoE` (#11211) · 3c06b673
  Cheng Wan authored Oct 07, 2025
  
  3c06b673
- [Feature] Add /tokenize and /detokenize OpenAI compatible endpoints (#9545) · 7c3f07db
  Adarsh Shirawalmath authored Oct 08, 2025
  
  7c3f07db
- [router][grpc] Refactor chat handler in grpc/ to use centralized orchestrator (#11314) · edd86b88
  Chang Su authored Oct 07, 2025
```
Co-authored-by: Simo Lin <linsimo.mark@gmail.com>
```
  edd86b88
- Rename `ngram_utils` -> `ngram_info` (#11316) · 4b4dc132
  Liangsheng Yin authored Oct 08, 2025
  
  4b4dc132
- Optimize copy_kv_cache for spec decoding (#11126) · 5a9170d9
  YAMY authored Oct 08, 2025
```
Co-authored-by: Xinyuan Tong <115166877+JustinTong0323@users.noreply.github.com>
```
  5a9170d9
- update sampling_params documentation with defaults (#11315) · c4d77774
  Xinyuan Tong authored Oct 07, 2025
  
  c4d77774
- [Chore] Update xgrammar 0.1.24 -> 0.1.25 (#10710) · 832c84fb
  DarkSharpness authored Oct 08, 2025
  
  832c84fb
- ci: unify the model launch method of nightly ci (#11230) · 64d1505c
  Mick authored Oct 08, 2025
  
  64d1505c
- Clean match_prefix and prepare_for_extend for mem cache V2 (#11200) · f3764c26
  cctry authored Oct 07, 2025
  
  f3764c26
- [oai serving chat] Add argument `--sampling-defaults` and fix... · 7ba3de0e
  Chang Su authored Oct 07, 2025
```
[oai serving chat] Add argument `--sampling-defaults` and fix `ChatCompletionRequest` defaults (#11304)
```
  7ba3de0e
07 Oct, 2025 10 commits
- [router] cleanup worker health check to return early (#11310) · fde9b963
  Simo Lin authored Oct 07, 2025
  
  fde9b963
- [router][grpc] Fix request_id extraction when n > 1 (#11311) · f094e0a4
  Chang Su authored Oct 07, 2025
  
  f094e0a4
- [router] support Openai router conversation API CRUD (#11297) · 4ed67c27
  Keyang Ru authored Oct 07, 2025
  
  4ed67c27
- [quantization] Properly ignore quantization for layers excluded in quant_config (#11205) · cd4b39a9
  Bowen Bao authored Oct 07, 2025
  
  cd4b39a9
- [router][grpc] Fix error message format in grpc chat handler (#11307) · 420c99ac
  Chang Su authored Oct 07, 2025
  
  420c99ac
- Update tool parser and related documentation (#11223) · e3c7f091
  Xinyuan Tong authored Oct 07, 2025
  
  e3c7f091
- [router][grpc] Fix sampling_params.stop_strs is None (#11306) · 6f1e03a4
  Chang Su authored Oct 07, 2025
  
  6f1e03a4
- [router] fix grpc connection conversion and add optimization (#11305) · f4affd4d
  Simo Lin authored Oct 07, 2025
  
  f4affd4d
- [Doc]: Best Practice for HICache (#11001) · df08bf9b
  hzh0425 authored Oct 08, 2025
```
Co-authored-by: 晟海 <huangtingwei.htw@antgroup.com>
Co-authored-by: huangtingwei <141888744+huangtingwei9988@users.noreply.github.com>
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
```
  df08bf9b
- [Doc] HiCache Design Documents (#11027) · 69efdd27
  ykwd authored Oct 08, 2025
```
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
```
  69efdd27