Commits · be1a85b7a2929f25c93d469fdd733a3576609e70 · OpenDAS / vllm_cscc · GitLab

26 Mar, 2026 25 commits
- Revert "[MoE Kernel] Flashinfer nvfp4 cutedsl moe kernel integration" (#38050) (#38169) · be1a85b7
  Zhewen Li authored Mar 26, 2026
```
Co-authored-by: Zhewen Li <zhewenli@inferact.ai>
```
  be1a85b7
- [Renderer] Consolidate factory methods (#38218) · 2e225f7b
  Cyrus Leung authored Mar 26, 2026
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  2e225f7b
- [bug-fix] GLM OCR Patch Merger context_dim (#37962) · 757eafcf
  Jared Wen authored Mar 26, 2026
```
Signed-off-by: JaredforReal <w13431838023@gmail.com>
```
  757eafcf
- [CI] Reorganize scoring tests (#38207) · dcdc1458
  wang.yuqi authored Mar 26, 2026
```
Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
```
  dcdc1458
- [ROCm][CI] Fix flaky GPTQ compile correctness test (#38161) · f2d16207
  Andreas Karatzas authored Mar 26, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  f2d16207
- [ROCm][CI] Fix wvSplitKrc mock argument order in test_rocm_unquantized_gemm (#38167) · 37a83007
  Andreas Karatzas authored Mar 26, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  37a83007
- [Refactor] Remove unused utils (#38153) · bf5eec63
  Wentao Ye authored Mar 26, 2026
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  bf5eec63
- DOC: Documentation pages fixes (#38125) · b1cb1d3d
  Mateusz Sokół authored Mar 26, 2026
```
Signed-off-by: Mateusz Sokół <mat646@gmail.com>
```
  b1cb1d3d
- [XPU] Disable xpu graph by default (#38193) · 6ae8bbd0
  Kunshang Ji authored Mar 26, 2026
```
Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>
```
  6ae8bbd0
- [Doc] Fix outdated reference to CUDAGraphManager (#38209) · a9213c0f
  Cyrus Leung authored Mar 26, 2026
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  a9213c0f
- [Model] Use helper function to run MM processors with token inputs (where applicable) (#38018) · 502c41a8
  Cyrus Leung authored Mar 26, 2026
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  502c41a8
- [Bugfix] Fix DeepGemm E8M0 accuracy degradation for Qwen3.5 FP8 on Blackwell (#38083) · 52069012
  Vadim Gimpelson authored Mar 26, 2026
```
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
```
  52069012
- [cpu][ci] remove soft-fail for Arm CI and add quant model tests (#37691) · 71161e8b
  Fadi Arafeh authored Mar 26, 2026
```
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
```
  71161e8b
- [Model] Add torch.compile support for InternVL vision encoder (#38049) · 38de8223
  Terry Gao authored Mar 25, 2026
```
Signed-off-by: tianrengao <terrygao87@gmail.com>
```
  38de8223
- [Bugfix] Fix benchmark_fused_collective.py (#38082) · 2bfbdca2
  Jee Jee Li authored Mar 26, 2026
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  2bfbdca2
- Add `/v1/chat/completions/batch` endpoint for batched chat completions (#38011) · 29080945
  Matej Rojec authored Mar 26, 2026
```
Signed-off-by: Matej Rojec <64556640+MatejRojec@users.noreply.github.com>
```
  29080945
- [Bugfix][CI] Fix Marlin FP8 Linear Kernel for Compressed Tensors Format (#38092) · e6bf9f15
  BadrBasowid authored Mar 26, 2026
```
Signed-off-by: BadrBasowid <Badr.Basowid@gmail.com>
Signed-off-by: BadrBasowid <61441185+BadrBasowid@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
```
  e6bf9f15
- Relocate Encoder CUDA graph manager (#38116) · 144030c8
  Woosuk Kwon authored Mar 25, 2026
```
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
Signed-off-by: Nick Hill <nickhill123@gmail.com>
Co-authored-by: Nick Hill <nickhill123@gmail.com>
```
  144030c8
- [Tool Parser][1/3] Pass tools to ToolParser constructor (#38029) · e2db2b42
  Flora Feng authored Mar 25, 2026
```
Signed-off-by: sfeng33 <4florafeng@gmail.com>
```
  e2db2b42
- [Revert] Remove DeepGEMM availability check in DeepseekV32IndexerMetadataBuilder (#38076) · 87f05d68
  Chauncey authored Mar 26, 2026
```
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
```
  87f05d68
- [Misc] Optimized check to encapsulate both CUDA and ROCm platforms (#34549) · 36f6aede
  Andreas Karatzas authored Mar 25, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  36f6aede
- Disable dual stream execution of input projection for Qwen3 (#38152) · 9704a5c3
  Xin Yang authored Mar 25, 2026
```
Signed-off-by: Xin Yang <xyangx@amazon.com>
```
  9704a5c3
- Fix minimax m2.5 nvfp4 kv scales weight loading (#37214) · 74056039
  Wei Zhao authored Mar 25, 2026
```
Signed-off-by: wzhao18 <wzhao18.sz@gmail.com>
```
  74056039
- [Bugfix] Fix Qwen3.5-FP8 Weight Loading Error on TPU (#37348) · d7d51a7e
  Jacob Platin authored Mar 25, 2026
```
Signed-off-by: Jacob Platin <jacobplatin@google.com>
```
  d7d51a7e
- Various Transformers v5 fixes (#38127) · 3c3c0842
  Harry Mellor authored Mar 26, 2026
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  3c3c0842
25 Mar, 2026 15 commits
- [Cohere] Enable Cohere-Transcribe (#38120) · 7b54f60d
  Ekagra Ranjan authored Mar 25, 2026
```
Signed-off-by: Ekagra Ranjan <3116519+ekagra-ranjan@users.noreply.github.com>
```
  7b54f60d
- [ROCm]: Update rope+kvcache fusion conditions and disable custom op by default (#36716) · a0e8c740
  Rohan Potdar authored Mar 25, 2026
```
Signed-off-by: Rohan138 <rohanpotdar138@gmail.com>
```
  a0e8c740
- [MultiModal] add support for numpy array embeddings (#38119) · 70a21528
  Guillaume Guy authored Mar 25, 2026
```
Signed-off-by: guillaume_guy <guillaume.guy@airbnb.com>
Signed-off-by: Guillaume Guy <guillaume.c.guy@gmail.com>
Co-authored-by: guillaume_guy <guillaume.guy@airbnb.com>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
```
  70a21528
- [ROCm] Utilize persistent MLA kernel from AITER (#36574) · 978fc18b
  Sathish Sanjeevi authored Mar 25, 2026
```
Signed-off-by: Sathish Sanjeevi <sathish.krishnan.p.s@gmail.com>
```
  978fc18b
- [ROCm] Fix MoE kernel test failures on gfx950 (#37833) · 7d6917be
  Andreas Karatzas authored Mar 25, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
Signed-off-by: Matthew Wong <Matthew.Wong2@amd.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Co-authored-by: Matthew Wong <Matthew.Wong2@amd.com>
```
  7d6917be
- [Core][KV Connector] Remove use of num_cached_tokens in error handling (#38096) · e38817fa
  Mark McLoughlin authored Mar 25, 2026
```
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
```
  e38817fa
- [Frontend] Move APIServerProcessManager target server fn (#38115) · 72cad44d
  Nick Hill authored Mar 25, 2026
```
Signed-off-by: Nick Hill <nickhill123@gmail.com>
```
  72cad44d
- [Misc] Reorganize inputs (#35182) · ba2f0acc
  Cyrus Leung authored Mar 26, 2026
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  ba2f0acc
- [MoE Kernel] Flashinfer nvfp4 cutedsl moe kernel integration (#38050) · 678b3c99
  Yongye Zhu authored Mar 25, 2026
  
  678b3c99
- [2/n] Migrate per_token_group_quant to torch stable ABI (#36058) · bf4cc9ed
  mikaylagawarecki authored Mar 25, 2026
```
Signed-off-by: Mikayla Gawarecki <mikaylagawarecki@gmail.com>
```
  bf4cc9ed
- [CI/Docs] Improve aarch64/DGX Spark support for dev setup (#38057) · 1ac2ef2e
  Ben Browning authored Mar 25, 2026
```
Signed-off-by: Ben Browning <bbrownin@redhat.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  1ac2ef2e
- [compile] Add some more startup tests for top models (#38046) · 6e37c46b
  Richard Zou authored Mar 25, 2026
```
Signed-off-by: Richard Zou <zou3519@gmail.com>
```
  6e37c46b
- [Refactor] Rename `WAITING_FOR_FSM` to `WAITING_FOR_STRUCTURED_OUTPUT_GRAMMAR` (#38048) · 1bf2ddd0
  Wentao Ye authored Mar 25, 2026
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  1bf2ddd0
- [Kernel] Optimize SM120 CUTLASS blockwise FP8 GEMM (#37970) · e7221180
  Necofish authored Mar 25, 2026
```
Signed-off-by: Necofish <liuxiangyang@mail.ustc.edu.cn>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
```
  e7221180
- [Bugfix] Preserve CUDA arch suffix (a/f) for SM12x — fixes NVFP4 NaN on desktop Blackwell (#37725) · 4a76ad12
  RobTand authored Mar 25, 2026
```
Signed-off-by: Rob Tand <robert.tand@icloud.com>
Co-authored-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
```
  4a76ad12