Commits · fc7980dbdae7a2eb91b680a15f2e3e44059e3ecd · OpenDAS / vllm_cscc

04 Feb, 2026 3 commits

[BugFix][Spec Decoding] Fix negative accepted tokens metric crash (#33729) · 1892993b
Nick Hill authored Feb 03, 2026
```
Signed-off-by: Nick Hill <nickhill123@gmail.com>
```
1892993b
cherry pick · 7d98f09b
Michael Goin authored Feb 03, 2026
```
Signed-off-by: Robert Shaw <rshaw@neuralmagic.com>
```
7d98f09b

[Bugfix] Disable RoutingMethodType.[Renormalize,RenormalizeNaive] TRTLLM... · daa2784b

Michael Goin authored Feb 03, 2026


[Bugfix] Disable RoutingMethodType.[Renormalize,RenormalizeNaive] TRTLLM per-tensor FP8 MoE (#33620)
Signed-off-by: mgoin <mgoin64@gmail.com>
(cherry picked from commit e346e2d0

)
Signed-off-by: Robert Shaw <rshaw@neuralmagic.com>

daa2784b

03 Feb, 2026 4 commits

[torch.compile] Don't do the fast moe cold start optimization if there is... · e4bf6ed9

Richard Zou authored Feb 02, 2026


[torch.compile] Don't do the fast moe cold start optimization if there is speculative decoding (#33624)
Signed-off-by: Richard Zou <zou3519@gmail.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
(cherry picked from commit 5eac9a1b)

e4bf6ed9

[torch.compile] Speed up MOE handling in forward_context (#33184) · 611b1875
Richard Zou authored Jan 27, 2026
```
Signed-off-by: Richard Zou <zou3519@gmail.com>
(cherry picked from commit d9aa39a3)
```
611b1875
[Misc][Build] Lazy load cv2 in nemotron_parse.py (#33189) · eec3546b
Kiersten Stokes authored Jan 29, 2026
```
Signed-off-by: kiersten-stokes <kierstenstokes@gmail.com>
(cherry picked from commit 9e138cb0)
```
eec3546b

[Release] Fix format and cherry-pick (#33618) · 31a64c63

Zhewen Li authored Feb 02, 2026


Signed-off-by: zhewenli <zhewen@inferact.ai>
Co-authored-by: zhewenli <zhewen@inferact.ai>

31a64c63

02 Feb, 2026 10 commits

[Release] patch step3p5 attention class in v0.15.1 release (#33602) · 57eae2f8
Zhewen Li authored Feb 02, 2026
```
Signed-off-by: zhewenli <zhewen@inferact.ai>
Co-authored-by: zhewenli <zhewen@inferact.ai>
```
57eae2f8
[Fix] prefix cache hit rate == 0 bug with gpt-oss style models (#33524) · f0d00586
Yifan Qiao authored Feb 01, 2026
```
Signed-off-by: Yifan Qiao <yifanqiao@berkeley.edu>
(cherry picked from commit a01ef3fa)
```
f0d00586

[Models] Step-3.5-Flash (#33523) · 8b45c58f

csy0225 authored Feb 02, 2026


Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
Co-authored-by: i-zhangmingming <i-zhangmingming@stepfun.com>
Co-authored-by: xiewuxun <xiewuxun@stepfun.com>
Co-authored-by: zetaohong <i-hongzetao@stepfun.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
(cherry picked from commit c3b40dc3)

8b45c58f

fix: Add SM120 (RTX Blackwell) support for FlashInfer CUTLASS NVFP4 MoE kernels (#33417) · 15ebd0ce
René Honig authored Jan 31, 2026
```
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
(cherry picked from commit 07978117)
```
15ebd0ce

[fix][torch.compile] Fix cold-start compilation time increase by adding kv... · 29152683

Luka Govedič authored Jan 31, 2026


[fix][torch.compile] Fix cold-start compilation time increase by adding kv cache update to splitting ops (#33441)
Signed-off-by: Luka Govedič <lgovedic@redhat.com>
Co-authored-by: Richard Zou <zou3519@gmail.com>
(cherry picked from commit 15f40b20)

29152683

[BugFix] Fix whisper FA2 + full cudagraphs (#33360) · d984d664

Lucas Wilkinson authored Jan 30, 2026


Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
(cherry picked from commit 0a3c71e7)

d984d664

[Bugfix][ROCm] Fixing the skinny gemm dispatch logic from #32831 (#33366) · 5f45b0b7
Gregory Shtrasberg authored Jan 30, 2026
```
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
(cherry picked from commit 31aedfe7)
```
5f45b0b7
[Bugfix] Enable Triton MoE for FP8 per-tensor dynamic (#33300) · 6ff16b77
Michael Goin authored Jan 29, 2026
```
Signed-off-by: mgoin <mgoin64@gmail.com>
(cherry picked from commit bfb9bdaf)
```
6ff16b77

[Bugfix] Fix Qwen3-VL-Reranker load. (#33298) · 1ed963d4

wang.yuqi authored Jan 29, 2026


Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
Signed-off-by: wang.yuqi <noooop@126.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
(cherry picked from commit abb34ac4)

1ed963d4

[Bugfix] Register fp8 cutlass_group_gemm as supported for only SM90+SM100 (#33285) · 39e8b493
Michael Goin authored Jan 28, 2026
```
Signed-off-by: mgoin <mgoin64@gmail.com>
(cherry picked from commit 1bd47d6e)
```
39e8b493

28 Jan, 2026 5 commits

Revert "Enable Cross layers KV cache layout at NIXL Connector (#30207)" (#33241) · fe18ce4d

Or Ozeri authored Jan 28, 2026


Signed-off-by: Or Ozeri <oro@il.ibm.com>
Co-authored-by: Kevin H. Luu <khluu000@gmail.com>
(cherry picked from commit 2e8de867)

fe18ce4d

[BugFix] Fix P/D with non-MoE DP (#33037) · 7779de34

Nick Hill authored Jan 27, 2026


Signed-off-by: Nick Hill <nickhill123@gmail.com>
(cherry picked from commit 0cd259b2)

7779de34

[Bugfix] Fix DeepseekV32 `AssertionError: num_kv_heads == 1` (#33090) · 0d8ce320
Nicolò Lucchesi authored Jan 27, 2026
```
Signed-off-by: NickLucche <nlucches@redhat.com>
(cherry picked from commit 492a7983)
```
0d8ce320
[Bugfix] Disable CG for Whisper+FA2 (#33164) · d51e1f8b
Nicolò Lucchesi authored Jan 27, 2026
```
Signed-off-by: NickLucche <nlucches@redhat.com>
(cherry picked from commit 1f3a2c29)
```
d51e1f8b

[Models] Kimi-K2.5 (#33131) · 5042815a

Roger Wang authored Jan 26, 2026


Signed-off-by: wanglinian <wanglinian@stu.pku.edu.cn>
Signed-off-by: wangln19 <96399074+wangln19@users.noreply.github.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: youkaichao <youkaichao@gmail.com>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: wanglinian <wanglinian@stu.pku.edu.cn>
Co-authored-by: wangln19 <96399074+wangln19@users.noreply.github.com>
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
(cherry picked from commit b539f988)

5042815a

27 Jan, 2026 10 commits

[Models] Kimi-K2.5 (#33131) · b539f988

Roger Wang authored Jan 26, 2026


Signed-off-by: wanglinian <wanglinian@stu.pku.edu.cn>
Signed-off-by: wangln19 <96399074+wangln19@users.noreply.github.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: youkaichao <youkaichao@gmail.com>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: wanglinian <wanglinian@stu.pku.edu.cn>
Co-authored-by: wangln19 <96399074+wangln19@users.noreply.github.com>
Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

b539f988

[code clean] remove duplicate code (#33135) · b781eeaa
Ning Xie authored Jan 27, 2026
```
Signed-off-by: Andy Xie <andy.xning@gmail.com>
```
b781eeaa
[Frontend] Cleanup serving engine (#33103) · e0b005d9
Cyrus Leung authored Jan 27, 2026
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
e0b005d9
[torch.compile] Stop assuming 32 bit indexing (#33113) · 3b8f0fe5
Richard Zou authored Jan 26, 2026
```
Signed-off-by: Richard Zou <zou3519@gmail.com>
```
3b8f0fe5
[Frontend] Reduce mixin usage in serving pooling (#33101) · c831911b
Cyrus Leung authored Jan 27, 2026
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
c831911b
[Perf] avoid duplicate mem_get_info() call in get_current_memory_usage (#33064) · 157caf51
Paco Xu authored Jan 27, 2026
```
Signed-off-by: Paco Xu <paco.xu@daocloud.io>
```
157caf51
Fix IndexError with encoder-decoder models when using Custom Paged Attention (#33112) · c568581f
Strahinja Stamenkovic authored Jan 27, 2026
```
Signed-off-by: sstamenk <strahinja.stamenkovic@amd.com>
```
c568581f

fix: preserve native tool call ID in multi-turn tool calling (#32768) · 2d705343

wangln19 authored Jan 27, 2026


Signed-off-by: wanglinian <wanglinian@stu.pku.edu.cn>
Signed-off-by: wangln19 <96399074+wangln19@users.noreply.github.com>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Isotr0py <2037008807@qq.com>

2d705343

[MoE Refactor] Integrate Naive Prepare Finalize into MK (#32567) · 5a93b916

Robert Shaw authored Jan 26, 2026


Signed-off-by: Robert Shaw <robshaw@redhat.com>
Signed-off-by: Amir Klein <203507526+amirkl94@users.noreply.github.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: amirkl94 <203507526+amirkl94@users.noreply.github.com>

5a93b916

[Model Runner V2] Remove UvaBufferPool for cpu->gpu copy (#33055) · 6d86fde0

Woosuk Kwon authored Jan 26, 2026


Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Co-authored-by: Nick Hill <nhill@redhat.com>

6d86fde0

26 Jan, 2026 8 commits

[Bugfix][TPU] Return a Default fp8 MoE Backend (#32908) · 510ed1e8

XiongfeiWei authored Jan 26, 2026


Signed-off-by: Xiongfei Wei <isaacwxf23@gmail.com>
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>

510ed1e8

[Bugfix][MXFP4] Call `trtllm_fp4_block_scale_moe` with kwargs (#33104) · 8caffd92
Pengchao Wang authored Jan 26, 2026
```
Signed-off-by: Pengchao Wang <wpc@fb.com>
```
8caffd92
[fix] CPUDNNLGEMMHandler pointer baked into inductor artifact (#32913) · 58a05b0c
dolpm authored Jan 26, 2026
```
Signed-off-by: dolpm <34420038+dolpm@users.noreply.github.com>
```
58a05b0c

[Logging] add `--disable-access-log-for-endpoints` CLI option (#30011) · 6ee7f18f

Jared Wen authored Jan 27, 2026



Add a new CLI option --disable-access-log-for-endpoints to suppress
uvicorn access logs for specified endpoints (e.g., /health, /metrics, /ping).

This addresses the need to reduce log noise in production environments
where health check endpoints are frequently polled by load balancers or
monitoring systems, generating excessive log entries that obscure
meaningful request logs.

Fixes #29982
Signed-off-by: JaredforReal <w13431838023@gmail.com>

6ee7f18f

[Refactor] Remove unused `_moe_permute` function (#33108) · 8f987883
Wentao Ye authored Jan 26, 2026
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
8f987883

[Bugfix] Fix Dtypes for Pynccl Wrapper (#33030) · cf1167e5

Robert Shaw authored Jan 26, 2026


Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
(cherry picked from commit 43a013c3)

cf1167e5

[Bugfix] Fix Dtypes for Pynccl Wrapper (#33030) · 43a013c3

Robert Shaw authored Jan 26, 2026


Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>

43a013c3

[Model] Bump transformers version for test registry (#33100) · c25dbee4
Cyrus Leung authored Jan 27, 2026
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
c25dbee4