Commits · 8f3d67b50ac41bb815bb4ea634d8fea9d56bf208 · OpenDAS / vllm_cscc

"vscode:/vscode.git/clone" did not exist on "6bb59e94bb6ac039a9c39ff026e2774116a5237e"

19 Aug, 2025 2 commits
- Revert "skip moe_fused_gate" · a4b438d0
  zhuwenwen authored Aug 19, 2025
```
This reverts commit 7a6ab87a.
```
  a4b438d0
- skip moe_fused_gate · 7a6ab87a
  zhuwenwen authored Aug 19, 2025
  
  7a6ab87a
18 Aug, 2025 8 commits
- fix issue from merge · 4f359a91
  zhuwenwen authored Aug 18, 2025
  
  4f359a91
- [fix]解决v1 deepseek cudagraph模式显存占用增长 · e85207b7
  zhuwenwen authored Aug 18, 2025
  
  e85207b7
- 修改flashmla的接口 · b909d6fc
  zhuwenwen authored Aug 18, 2025
  
  b909d6fc
- 增加marlin对cache13的支持，以及新增flash mla的kvcache fp8的支持 · 22a46529
  zhuwenwen authored Aug 18, 2025
  
  22a46529
- debug and fix erro · e70b0ea0
  zhuwenwen authored Aug 18, 2025
  
  e70b0ea0
- fix mtp accept rate issu · 1555157e
  zhuwenwen authored Aug 18, 2025
  
  1555157e
- add v1 engine + deepseek r1 mtp + zero-overhead scheduler · 22a95571
  zhuwenwen authored Aug 18, 2025
  
  22a95571
- [fix]修复mtp eager模式下显存占用增加问题 · ac4cc84e
  zhuwenwen authored Aug 18, 2025
  
  ac4cc84e
17 Aug, 2025 10 commits
- fix: gptq marlin weight loading failure (#23066) · 0fc8fa75
  Simon Mo authored Aug 17, 2025
  
  0fc8fa75
- [XPU] fix xpu to set cudagraph batch sizes (#23044) · 21e39436
  Calvin Chen authored Aug 18, 2025
```
Signed-off-by: calvin chen <wen.chen@dynamia.ai>
```
  21e39436
- [Misc] Convert use_structured_output property into constant (#23060) · 6d243efe
  Woosuk Kwon authored Aug 17, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  6d243efe
- [Misc] Remove dead return (#23061) · c55bc1db
  Woosuk Kwon authored Aug 17, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  c55bc1db
- [Bugfix] should use stack instead of concat (#22972) · fe0411fc
  947132885 authored Aug 17, 2025
```
Signed-off-by: 947132885 <947132885@qq.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  fe0411fc
- [Kernel] Add cuda kernel for gpt_oss activation (#22951) · 4d4061b6
  Jee Jee Li authored Aug 17, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  4d4061b6
- [Misc] method name typo fix (#23042) · 87f48623
  Ning Xie authored Aug 17, 2025
```
Signed-off-by: Andy Xie <andy.xning@gmail.com>
```
  87f48623
- [Refactor] Defer tensor data construction in MultiModalKwargs (#23030) · 5c32143b
  Cyrus Leung authored Aug 17, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  5c32143b
- [UX] Separate marlin moe config logic from triton moe (#23006) · 94096a47
  Michael Goin authored Aug 16, 2025
  
  94096a47
- [Bugfix] fix qwen3 moe fp8 accuracy issue (#23031) · a258ad8b
  Jinzhen Lin authored Aug 17, 2025
```
Signed-off-by: Jinzhen Lin <jinzhen.ljz@antgroup.com>
```
  a258ad8b
16 Aug, 2025 17 commits
- [V1] Logits processors extensibility (#19912) · bf7f470b
  afeldman-nm authored Aug 16, 2025
```
Signed-off-by: Andrew Feldman <afeldman@redhat.com>
Signed-off-by: Andrew Feldman <afeld2012@gmail.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: Nick Hill <nhill@redhat.com>
Co-authored-by: Nick Hill <nhill@redhat.com>
Co-authored-by: Andrew Feldman <afeld2012@gmail.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  bf7f470b
- [Kernel/Quant] Remove AQLM (#22943) · 4fc722ec
  Michael Goin authored Aug 16, 2025
```
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
```
  4fc722ec
- [Bugfix gpt-oss] Fix float32 convert for flashinfer sink support (#23016) · 000cceca
  Michael Goin authored Aug 16, 2025
```
Signed-off-by: mgoin <mgoin64@gmail.com>
```
  000cceca
- [Frontend] Added support for HermesToolParser for models without special tokens (#16890) · 68373d31
  Woonggi Min authored Aug 17, 2025
```
Signed-off-by: minpeter <kali2005611@gmail.com>
```
  68373d31
- Fix handling of `max_num_batched_tokens` for pooling tasks (#23004) · 52ce1420
  Maximilien de Bayser authored Aug 16, 2025
```
Signed-off-by: Max de Bayser <mbayser@br.ibm.com>
```
  52ce1420
- [New Model]mBART model (#22883) · 829bbd78
  汪志鹏 authored Aug 16, 2025
```
Signed-off-by: 汪志鹏 <wangzhipeng628@gmail.com>
```
  829bbd78
- [Refactor] Allow optional MultiModalKwargsItem in IPC (#23022) · 4dff91c9
  Cyrus Leung authored Aug 16, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  4dff91c9
- Add docs for PrefixRepetitionDataset + enable usage with `vllm bench throughput` (#23012) · de9cb617
  Seiji Eicher authored Aug 16, 2025
```
Signed-off-by: Seiji Eicher <seiji@anyscale.com>
Co-authored-by: Roger Wang <hey@rogerw.me>
```
  de9cb617
- [Core] Make cudagraph check cuda platform only (#23005) · 933f4533
  Chengji Yao authored Aug 16, 2025
```
Signed-off-by: Chengji Yao <chengjiyao@gmail.com>
Signed-off-by: Chengji Yao <chengjiyao@google.com>
Co-authored-by: Chengji Yao <chengjiyao@gmail.com>
Co-authored-by: Li, Jiang <jiang1.li@intel.com>
```
  933f4533
- [Multimodal] Update Tensor schema test to cover arbitrary shape mm inputs (#22867) · cc826a20
  Isotr0py authored Aug 16, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  cc826a20
- [BugFix] Add support for loading prompt embeds tensors serialized on... · 78863f8c
  Andrew Sansom authored Aug 16, 2025
```
[BugFix] Add support for loading prompt embeds tensors serialized on unavailable devices and sparse tensors (#22962)
Signed-off-by: Andrew Sansom <andrew@protopia.ai>
```
  78863f8c
- [XPU]avoid circular import during XPU init (#23017) · 7caec10e
  Kunshang Ji authored Aug 16, 2025
```
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
```
  7caec10e
- [V1] support min_tokens for detokener (#22014) · e4e37ded
  Calvin Chen authored Aug 16, 2025
```
Signed-off-by: calvin chen <wen.chen@dynamia.ai>
Co-authored-by: Nick Hill <nhill@redhat.com>
```
  e4e37ded
- [Frontend] Avoid list copies in `serving_chat.py` (#22947) · f6b50405
  Nick Hill authored Aug 15, 2025
```
Signed-off-by: Nick Hill <nhill@redhat.com>
```
  f6b50405
- [Bugfix] Fix DeepSeek MTP (#22934) · fbd88728
  Benjamin Chislett authored Aug 15, 2025
```
Signed-off-by: Benjamin Chislett <benjamin.chislett@centml.ai>
```
  fbd88728
- [Kernel] Simplify `get_kv_cache_layout` and cache `use_trtllm_attention` env-dependent bit (#22735) · 070da660
  Nicolò Lucchesi authored Aug 16, 2025
```
Signed-off-by: NickLucche <nlucches@redhat.com>
```
  070da660
- [Misc] Support passing multiple request ids at once to `AsyncLLM.abort()` (#22944) · ad0297d1
  Nick Hill authored Aug 15, 2025
```
Signed-off-by: Nick Hill <nhill@redhat.com>
```
  ad0297d1
15 Aug, 2025 3 commits
- [BugFix] Make `run_once` thread-safe (#22978) · 236b864e
  Yichen Yan authored Aug 16, 2025
```
Signed-off-by: <wenji.yyc@alibaba-inc.com>
Signed-off-by: Yichen Yan <wenji.yyc@alibaba-inc.com>
```
  236b864e
- Support multiple attention groups for KV sharing (#22672) · 3e2f7985
  Yong Hoon Shin authored Aug 15, 2025
```
Signed-off-by: Yong Hoon Shin <yhshin@meta.com>
```
  3e2f7985
- [v1] Move block_hashes from KVCacheManager to Request.block_hashes (#19728) · c280066f
  Or Ozeri authored Aug 16, 2025
```
Signed-off-by: Or Ozeri <oro@il.ibm.com>
```
  c280066f