Commits · 62d54ba46db25b95de2d21e46f4b57b5502ed747 · OpenDAS / vllm_cscc · GitLab

23 Nov, 2025 6 commits
- [Model Runner V2] Optimize CUDA graph capture time (#29275) · 62d54ba4
  Woosuk Kwon authored Nov 23, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  62d54ba4
- [Model Runner V2] Support spec decoding [1/N] (#29274) · b004c004
  Woosuk Kwon authored Nov 23, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  b004c004
- [Model Runner V2] Change bookkeeping logic in preparation for spec decoding (#29194) · 7f12c82f
  Woosuk Kwon authored Nov 23, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  7f12c82f
- [Bugfix] Use lazy string reference for DeepseekV3Config in config registry (#28958) · 6fb0215e
  Luke authored Nov 23, 2025
```
Signed-off-by: Luke <yq0536@gmail.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  6fb0215e
- [Model Runner V2] Minor fix for cudagraph_utils (#29256) · 20ee418a
  Woosuk Kwon authored Nov 22, 2025
  
  20ee418a
- [Fix] Add SM check to flashinfer MOE backend (#29144) · 5f96c00c
  jiahanc authored Nov 22, 2025
```
Signed-off-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
```
  5f96c00c
22 Nov, 2025 24 commits
- Refactor: Move CUDA graph dispatch logic earlier (#27382) · df78aeef
  Yizhou authored Nov 23, 2025
```
Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>
```
  df78aeef
- [BugFix] Fix chunked prompt logprobs + preemption (#29071) · 7df331c6
  Nick Hill authored Nov 22, 2025
  
  7df331c6
- [Bugfix] Use HF config fields as fallback when loading Mistral config (#29239) · d1cf8214
  Cyrus Leung authored Nov 23, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  d1cf8214
- [perf][cpu] Accelerate paged attention GEMMs (QK, PV) on Arm CPUs with NEON (#29193) · 730bd353
  Fadi Arafeh authored Nov 22, 2025
```
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
```
  730bd353
- chore: add RTX_PRO_6000 GLM4.6-FP8 kernel tuning (#29240) · f55c76c2
  Federico authored Nov 22, 2025
  
  f55c76c2
- Fix EVS crash when using `video_embeds` inputs in Qwen2.5-VL (#29232) · d84d8f44
  ZiTian Zhao authored Nov 22, 2025
```
Signed-off-by: zitian.zhao <zitian.zhao@tencentmusic.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
```
  d84d8f44
- [Misc] Fix pre-commit (#29238) · ae668183
  Cyrus Leung authored Nov 22, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  ae668183
- [BugFix] Fix returned logprobs with spec decode + prefill chunking (#29216) · d44a63c6
  Nick Hill authored Nov 22, 2025
```
Signed-off-by: Nick Hill <nhill@redhat.com>
```
  d44a63c6
- [Attention] Refactor FA `block_size` limitations to hybrid models only (#29084) · 066209a0
  Nicolò Lucchesi authored Nov 22, 2025
```
Signed-off-by: NickLucche <nlucches@redhat.com>
```
  066209a0
- [tiny] Remove unsupported TRITON_MLA backend from batch invariance (#28832) · 5f7209a7
  Bram Wasti authored Nov 22, 2025
```
Signed-off-by: Bram Wasti <bwasti@meta.com>
Signed-off-by: Bram Wasti <bwasti@fb.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
```
  5f7209a7
- Fix: Resolve circular import in model_loader/utils.py (#29189) · 6965a392
  Nandan Vallamdasu authored Nov 22, 2025
```
Signed-off-by: nandan2003 <nandan.vallamdasu@outlook.com>
Signed-off-by: Nandan Vallamdasu  <nandan.vallamdasu@outlook.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
```
  6965a392
- [Misc] Further clean up chunked prefill and prefix caching init (#29186) · 5a480258
  Cyrus Leung authored Nov 22, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  5a480258
- Handle triton kernel import exception (#29062) · 988ee66b
  jinghanhu authored Nov 22, 2025
  
  988ee66b
- [Frontend][Responses API] Multi-turn (with type: "output_text") support for... · ea38474a
  Mads Kildegård authored Nov 22, 2025
```
[Frontend][Responses API] Multi-turn (with type: "output_text") support for non-harmony requests (#29175)
Signed-off-by: Mads Kildegård <mkildegaard99@gmail.com>
```
  ea38474a
- [responsesAPI] parse reasoning item input (#28248) · 742e9ff6
  Andrew Xia authored Nov 21, 2025
```
Signed-off-by: Andrew Xia <axia@fb.com>
Co-authored-by: Andrew Xia <axia@fb.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
```
  742e9ff6
- [Model Runner V2] Limit cudagraph size to max decode batch size (#29221) · e9056056
  Woosuk Kwon authored Nov 21, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  e9056056
- [LoRA] Cleanup FusedMoEWithLoRA (#29187) · 1489902b
  Jee Jee Li authored Nov 22, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  1489902b
- [Bugfix]Fix a conditional to not check zero value (#28754) · 933f67ec
  Yanan Cao authored Nov 21, 2025
```
Signed-off-by: Yanan Cao <gmagogsfm@gmail.com>
```
  933f67ec
- [chore][LMCache connector] Remove useless logs from lmcache connector (#29069) · 77e1c035
  Yihua Cheng authored Nov 21, 2025
```
Signed-off-by: ApostaC <yihua98@uchicago.edu>
```
  77e1c035
- Add fused MoE config for H200 E160 N192 fp8 (#29182) · 052950e5
  FlintyLemming authored Nov 22, 2025
```
Signed-off-by: FlintyLemming <admin@flinty.moe>
```
  052950e5
- [Bugfix] Fix block size in block_table with PCP (#29094) · 5c8f2adf
  Jie Luo authored Nov 22, 2025
```
Signed-off-by: Livinfly <luojie3m@gmail.com>
```
  5c8f2adf
- [Model][Qwen3VL] Tune Triton w8a8 block fp8 kernel for L40s (#29217) · d045e22d
  Lukas Geiger authored Nov 22, 2025
```
Signed-off-by: Lukas Geiger <lukas.geiger94@gmail.com>
```
  d045e22d
- [CI] Bug: Fix triton import issue (#29202) · 1d34eb11
  Wentao Ye authored Nov 21, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  1d34eb11
- [BugFix] Fix Eagle `IndexError: list index out of range` for even `num_speculative_tokens` (#29102) · 30d64662
  Lucas Wilkinson authored Nov 21, 2025
```
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
```
  30d64662
21 Nov, 2025 10 commits
- [Model Runner V2] Optimize Gumbel Sampling Kernel (#29210) · e9af6ba6
  Woosuk Kwon authored Nov 21, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  e9af6ba6
- [KV Connector] Fix async connector prefix cache metrics (#28585) · c6fa3895
  Mark McLoughlin authored Nov 21, 2025
```
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Co-authored-by: Nicolò Lucchesi <nlucches@redhat.com>
```
  c6fa3895
- [BugFix] EPLB + B200 + DeepGEMM : Handle column-major scales tensor (#29162) · 3137991f
  Varun Sundar Rabindranath authored Nov 21, 2025
```
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
```
  3137991f
- Default model load/config/tokenizer to `mistral` format if relevant files exist (#28659) · 57430fc9
  Julien Denize authored Nov 21, 2025
```
Signed-off-by: Julien Denize <julien.denize@mistral.ai>
Signed-off-by: Julien Denize <40604584+juliendenize@users.noreply.github.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
```
  57430fc9
- [log] add weights loading time log to sharded_state loader (#28628) · 53a1ba6e
  Ning Xie authored Nov 22, 2025
```
Signed-off-by: Andy Xie <andy.xning@gmail.com>
```
  53a1ba6e
- [BugFix] Make sure to allocate worst case MoE workspace during profile run in... · 1840c5cb
  Lucas Wilkinson authored Nov 21, 2025
```
[BugFix] Make sure to allocate worst case MoE workspace during profile run in the DP + EP case (#27426)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
```
  1840c5cb
- [Chore] Fix pre-commit error after #25266 (#29190) · 1bed891f
  Woosuk Kwon authored Nov 21, 2025
  
  1bed891f
- [Deprecation] Deprecate `seed=None` (#29185) · ceca0605
  Cyrus Leung authored Nov 22, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  ceca0605
- [NIXL] Fix after virtual block_size for host_buffer with heter kv_layout (#29122) · 460d02a4
  Chendi.Xue authored Nov 21, 2025
```
Signed-off-by: Chendi Xue <chendi.xue@intel.com>
```
  460d02a4
- Add TRTLLM MoE NVFP4 kernel to CompressedTensorsW4A4MoeMethod (#28892) · b4c8fbaa
  Mingyuan Ma authored Nov 21, 2025
```
Signed-off-by: mingyuanm <mingyuanm@nvidia.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
```
  b4c8fbaa