Commits · 9fac6aa30b669de75d8718164cd99676d3530e7d · OpenDAS / vllm_cscc · GitLab

18 Sep, 2025 40 commits
- [BugFix] Fix DeepGEMM warmup, no m.weight_scale_inv (#25206) · 9fac6aa3
  Lucas Wilkinson authored Sep 18, 2025
```
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
```
  9fac6aa3
- [KV offload][1b/N] rename offloading to kv_offload (#25191) · a53ad626
  Or Ozeri authored Sep 18, 2025
```
Signed-off-by: Or Ozeri <oro@il.ibm.com>
```
  a53ad626
- [V0 Deprecation] Remove unused async_timeout.py (#25190) · 1c3dad22
  Woosuk Kwon authored Sep 18, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  1c3dad22
- [Bug] Fix torch Compilation Cache Hit Error (#25093) · d2a30a2d
  Wentao Ye authored Sep 18, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  d2a30a2d
- [Bug] Fix `returned_lse` not Defined issue (#25106) · 75fb112d
  Wentao Ye authored Sep 18, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
```
  75fb112d
- [feat]: Create interface for model-specific M-RoPE (#24194) · 38db529f
  Aziz authored Sep 18, 2025
```
Signed-off-by: AzizCode92 <azizbenothman76@gmail.com>
Signed-off-by: Aziz <azizbenothman76@gmail.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>
```
  38db529f
- [fix]: remove data type hardcoding from gptoss model implementation (#23807) · 064cac7b
  Nikhil Gupta authored Sep 18, 2025
```
Signed-off-by: Nikhil Gupta <nikhil.gupta2@arm.com>
```
  064cac7b
- [V0 Deprecation] Remove AsyncLLMEngine (#25025) · e19bce40
  Woosuk Kwon authored Sep 18, 2025
```
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  e19bce40
- [KV offload][1/N] Introduce an offloading component (#19848) · 505805b6
  Or Ozeri authored Sep 18, 2025
```
Signed-off-by: Or Ozeri <oro@il.ibm.com>
```
  505805b6
- [ROCm][AITER][Bugfix] Switch AITER to use PIECEWISE_AND_FULL compilation (#25104) · bbdc0f23
  Rohan Potdar authored Sep 18, 2025
```
Signed-off-by: Rohan138 <rohanpotdar138@gmail.com>
```
  bbdc0f23
- [spec decode] Fix MTP inference path for MiMo-7B model (#25136) · c4cb0af9
  qizixi authored Sep 18, 2025
```
Signed-off-by: zixi-qi <qizixi@meta.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
```
  c4cb0af9
- Enable Allgather/ReduceScatter backend for NaiveAllToAll (#23964) · 2ea50e97
  Shu Wang authored Sep 18, 2025
```
Signed-off-by: Shu Wang. <shuw@nvidia.com>
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Signed-off-by: Shu Wang <shuw@nvidia.com>
Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
```
  2ea50e97
- [Docs] Fix warnings in mkdocs build (continued) (#25163) · b419937c
  Hyogeun Oh (오효근) authored Sep 19, 2025
```
Signed-off-by: Zerohertz <ohg3417@gmail.com>
```
  b419937c
- [New Model] Support BertForTokenClassification / Named Entity Recognition (NER) task (#24872) · 5f696c33
  wang.yuqi authored Sep 18, 2025
```
Signed-off-by: wang.yuqi <noooop@126.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  5f696c33
- feat(api): Return 503 on /health when engine is dead (#24897) · 67244c86
  dongbo910220 authored Sep 18, 2025
```
Signed-off-by: dongbo910220 <1275604947@qq.com>
Co-authored-by: Claude <noreply@anthropic.com>
```
  67244c86
- [PERF] Add `conv1d` metadata to GDN attn (#25105) · 072d7e53
  Vadim Gimpelson authored Sep 18, 2025
```
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
```
  072d7e53
- [Kernel] Decouple Tile Size from Block Size in Triton Unified Attention Kernel (#21197) · 01a583fe
  jvlunteren authored Sep 18, 2025
```
Signed-off-by: Jan van Lunteren <jvl@zurich.ibm.com>
```
  01a583fe
- Refactor dense FP8 tensor/channel/block utils and add CT FP8 block (#21404) · fbd6523a
  Michael Goin authored Sep 18, 2025
  
  fbd6523a
- [Structured Output][Refactor] Move `apply_grammar_bitmask()` method from... · 470484a4
  Shanshan Shen authored Sep 18, 2025
```
[Structured Output][Refactor] Move `apply_grammar_bitmask()` method from `ModelRunner` to structured output utils (#21999)
Signed-off-by: shen-shanshan <467638484@qq.com>
```
  470484a4
- [Misc] Clean up flags in `vllm bench serve` (#25138) · 21da7334
  Roger Wang authored Sep 18, 2025
```
Signed-off-by: Roger Wang <hey@rogerw.io>
```
  21da7334
- [Bugfix][Mamba] - Fix Conv State Kernel FP32 Support (#24883) · 66072b36
  Asaf Joseph Gardin authored Sep 18, 2025
```
Signed-off-by: asafg <39553475+Josephasafg@users.noreply.github.com>
```
  66072b36
- Fix `validate-config` pre-commit check (#25157) · 3ed1ec4a
  Harry Mellor authored Sep 18, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  3ed1ec4a
- Fix forward reference warning in documentation (#25150) · 5a33ae9a
  Harry Mellor authored Sep 18, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  5a33ae9a
- [Docs] add the parallel sampling usage in LLMEngine and AsyncLLM (#24222) · c9ff9e6f
  William Song authored Sep 18, 2025
  
  c9ff9e6f
- [Docs] Fix pooling-params doc references in openai_compatible_server.md (#24939) · eaffe448
  Kay Yan authored Sep 18, 2025
  
  eaffe448
- Move `StructuredOutputsConfig` from `config/__init__.py` to `config/structured_outputs.py` (#25153) · 8ed039d5
  Harry Mellor authored Sep 18, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  8ed039d5
- [Model] Improve Pooling Model (#25149) · 37970105
  Jee Jee Li authored Sep 18, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  37970105
- [Frontend] Support setting logprobs to -1 (#25031) · cc935fdd
  Chauncey authored Sep 18, 2025
```
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
```
  cc935fdd
- [Chore] Cleanup guided namespace, move to structured outputs config (#22772) · 29283e89
  Aaron Pham authored Sep 18, 2025
```
Signed-off-by: Aaron Pham <contact@aarnphm.xyz>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  29283e89
- [Doc] Fix cross-reference warnings (#25058) · 05b044e6
  Punitvara authored Sep 18, 2025
```
Signed-off-by: Punit Vara <punitvara@gmail.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  05b044e6
- [Qwen] Add fp8 checkpoint support for qwen3-next. (#25079) · ef7eefe1
  Tao He authored Sep 18, 2025
```
Signed-off-by: Tao He <linzhu.ht@alibaba-inc.com>
```
  ef7eefe1
- [Bugfix] when use s3 model cannot use default load_format (#24435) · 350c94de
  rongfu.leng authored Sep 18, 2025
```
Signed-off-by: rongfu.leng <rongfu.leng@daocloud.io>
Co-authored-by: 22quinn <33176974+22quinn@users.noreply.github.com>
```
  350c94de
- Retrieve `sliding_window` from text config in Gemma3 MM (#25085) · f4cd80f9
  Harry Mellor authored Sep 18, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  f4cd80f9
- [CLI] Use streaming in CLI chat and completion commands (#23769) · e111d5b0
  Simon Mo authored Sep 17, 2025
```
Signed-off-by: simon-mo <simon.mo@hey.com>
```
  e111d5b0
- [benchmark] add peak throughput metrics and plot (#23867) · a904ea78
  Simon Mo authored Sep 17, 2025
```
Signed-off-by: simon-mo <simon.mo@hey.com>
```
  a904ea78
- [Spec Decode] Efficient padded speculation (#24539) · b7433ca1
  Benjamin Chislett authored Sep 18, 2025
```
Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
```
  b7433ca1
- [EPLB] Add EPLB support for hunyuan_v1 (#23078) · 9d8a2d86
  YiwenC authored Sep 17, 2025
  
  9d8a2d86
- [XPU] Whisper model support on XPU Platform (#25123) · 3bc18127
  Chaojun Zhang authored Sep 18, 2025
```
Signed-off-by: chzhang <chaojun.zhang@intel.com>
```
  3bc18127
- Mark prompt logprobs as incompatible with prompt embeds at API level (#25077) · bec060fd
  Andrew Sansom authored Sep 17, 2025
```
Signed-off-by: Andrew Sansom <andrew@protopia.ai>
```
  bec060fd
- [Model] enable data parallel for InternVL vision encoder (#23909) · 52bc9d5b
  YiwenC authored Sep 17, 2025
```
Signed-off-by: Yiwen Chen <yiwen66@berkeley.edu>
Signed-off-by: YiwenC <54658925+666even666@users.noreply.github.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
```
  52bc9d5b