Commit Graph

Select branches

Hide Pull Requests

main

patches

sm86

Mono Color

• 3952731e8f [New Model]: Support Qwen3 Embedding & Reranker (#19260) wang.yuqi 2025-06-11 11:07:30 +08:00
• 77f0d465d0 [BugFix] Allow use_cudagraph to work with dynamic VLLM_USE_V1 (#19390) Richard Zou 2025-06-10 19:54:41 -04:00
• 22c3c0aa4a Add H20-3e fused MoE kernel tuning configs for Qwen3-235B-A22B-FP8 (#19401) Xu Wenqing 2025-06-11 07:23:57 +08:00
• 33f8dba7c6 [Model] use AutoWeightsLoader for commandr (#19399) py-andy-c 2025-06-10 15:42:21 -07:00
• 5241ca50d6 [ROCm][V1] Adding ROCm to the list of plaforms using V1 by default (#19440) Gregory Shtrasberg 2025-06-10 18:06:15 -04:00
• da9b523ce1 [Docs] Note that alternative structured output backends are supported (#19426) Russell Bryant 2025-06-10 12:20:00 -04:00
• b6553be1bc [Misc] Slight improvement of the BNB (#19418) 3 people 2025-06-10 21:51:49 +08:00
• 64a9af5afa Simplify ep kernels installation (#19412) youkaichao 2025-06-10 20:06:08 +08:00
• e4248849ec [BugFix][CPU] Fix CPU CI by ignore collecting test_pixtral (#19411) Li, Jiang 2025-06-10 20:02:40 +08:00
• 467bef18a3 [BugFix][FlashInfer] Fix attention backend interface mismatch with unexpected keyword use_irope (#19134) Rachel Guo 2025-06-10 01:48:51 -07:00
• 5f1ac1e1d1 Revert "[v1] Add fp32 support to v1 engine through flex attn" (#19404) Isotr0py 2025-06-10 16:30:20 +08:00
• 9368cc90b2 Automatically bind CPU OMP Threads of a rank to CPU ids of a NUMA node. (#17930) Louie TsaiandLi, Jiang <bigpyj64@gmail.com> 2025-06-09 23:22:05 -07:00
• 32b3946bb4 Add clear documentation around the impact of debugging flag (#19369) Anna Pendleton 2025-06-09 23:16:09 -07:00
• 6b1391ca7e [Misc] refactor neuron_multimodal and profiling (#19397) Reidandreidliu41 2025-06-10 14:12:42 +08:00
• a3f66e75d1 Add security warning to bug report template (#19365) Russell BryantandCopilot 2025-06-10 02:06:36 -04:00
• 319cb1e351 [Core] Batch multi modal input using pinned memory (#19169) Lukas Geiger 2025-06-10 07:44:59 +02:00
• 1efef71645 [Bugfix] Fix modelscope token passed in (#19389) Li WangandJee Jee Li 2025-06-10 13:39:37 +08:00
• 646d62f636 [Core] Use tuple for kv cache group block ids (#19175) Nick Hill 2025-06-09 22:01:17 -07:00
• 6cd4ae8acd [Frontend] Add tqdm_leave_pbar to control progress bar visibility (#19357) Reidandreidliu41 2025-06-10 12:55:09 +08:00
• c016047ed7 Fix docs/mkdocs/hooks/remove_announcement.py (#19382) Harry Mellor 2025-06-10 05:36:54 +01:00
• 9af6d22e4c Use xla flag to improve the quantized model performance (#19303) XiongfeiWei 2025-06-09 18:28:45 -07:00
• 4589b94032 [Bugfix] Fix benchmark_moe.py (#19016) Tianyu Guo 2025-06-10 09:04:36 +08:00
• cc867be19c [V1] Reuse V0's memory_profiling util for gpu worker memory profiling (#19312) Ye (Charlotte) Qi 2025-06-09 17:40:01 -07:00
• 3a7cd627a8 [Misc] Fix a config typo in disable_hybrid_kv_cache_manager configuration (#19383) Siyuan Liu 2025-06-09 16:41:51 -07:00
• 8058c91108 [HOT-FIX] Add kv_sharing_target_layer_name argument to cutlass_mla backend (#19374) Pavani Majety 2025-06-09 16:00:07 -07:00
• 7d44c469fe [TPU]Fix KV cache sharing tests (#19371) Siyuan Liu 2025-06-09 15:38:15 -07:00
• 31f58be96a [Frontend] Make TIMEOUT_KEEP_ALIVE configurable through env var (#18472) liusiqian-tal 2025-06-10 05:41:21 +08:00
• ebb2f383b8 [Quantization] Bump compressed-tensors version (#19295) Kyle Sayers 2025-06-09 17:33:15 -04:00
• c1c7dbbeeb [Bugfix][Core] Prevent token lengths exceeding max_model_len in V0 (#19348) 22quinn 2025-06-09 08:01:29 -07:00
• 5cf2daea9a [Misc] Fixes and Optimizations for DeepEP + DeepGEMM combination. (#19298) Varun Sundar RabindranathandVarun 2025-06-09 10:50:39 -04:00
• b8089195b4 [v1] Add fp32 support to v1 engine through flex attn (#19319) Isotr0py 2025-06-09 22:10:44 +08:00
• 770e5dcdb8 [full_graph] Fix query_start_loc padding (#19321) Yinghai Lu 2025-06-09 06:32:56 -07:00
• c57c9415b1 [Docs] Fix a bullet list in usage/security.md (#19358) Michael Yao 2025-06-09 21:28:51 +08:00
• 01810f9236 [CI] Introduce rules for llama auto-label (#19323) Lu Fang 2025-06-09 20:05:42 +08:00
• 59abbd84f9 [Fix] Allow kernel compilation for CUDA capability 8.7 (#19328) Conroy Cheers 2025-06-09 19:57:23 +10:00
• 95a6568b5c [CI/Build] Fix LoRA test (#19350) Jee Jee Li 2025-06-09 17:52:10 +08:00
• 0eca5eacd0 [Doc] Fix description in the Automatic Prefix Caching design doc (#19333) Se7en 2025-06-09 17:30:02 +08:00
• 12e5829221 [doc] improve ci doc (#19307) Reidandreidliu41 2025-06-09 15:26:12 +08:00
• 3a4d417707 [Misc] Cleanup compilation tests (#19343) Richard Zou 2025-06-09 03:05:44 -04:00
• 8335667c22 [Frontend] Remove unreachable code from llm.py (#19288) Kseniya Parkhamchuk 2025-06-08 21:22:10 -05:00
• e1c4380d4c [Misc] Add documentation update reminder to PR template (#19289) Isotr0py 2025-06-09 10:20:53 +08:00
• e31ae3de36 [Deprecation] Remove inputs arg fallback in Engine classes (#18799) Cyrus Leung 2025-06-09 10:19:56 +08:00
• 2ffb9b6e07 [Bugfix] model_max_length should consider max_model_len in tokenizer_config (#19201) wang.yuqi 2025-06-08 22:17:53 +08:00
• cda10fa3e2 [Multi Modal] Add an env var for message queue max chunk bytes (#19242) jennyyyyzhenandyZhen 2025-06-08 06:39:12 -07:00
• c123bc33f9 [Quantization] Add compressed-tensors NVFP4 support (#18312) Dipika Sikka 2025-06-08 06:05:55 -07:00
• b9a1791e2c [Hardware][POWER] Add IBM POWER11 Support to CPU Extension Detection (#19082) Akash kaothalkarandAkash Kaothalkar 2025-06-08 14:47:14 +05:30
• 989dcee981 Add H20-3e fused MoE kernel tuning configs for Qwen3-235B-A22B (#19315) Xu Wenqing 2025-06-08 16:07:02 +08:00
• 3d64d366e0 [Misc] Change tests/compile to use VLLM_V1 by default (#19302) Richard Zou 2025-06-08 04:06:48 -04:00
• eaa2e51088 [Bugfix] Re-enable use_cudagraph in vLLM v1 (#19299) Richard Zou 2025-06-07 20:56:12 -04:00
• d77f7fb871 [Bugfix]: Fix TypeError: 'float' object cannot be interpreted as an integer (#19283) Chauncey 2025-06-08 08:16:31 +08:00
• 2d8476e465 [BugFix][V1] Fix memory profiling bug (#18974) Luka Govedič 2025-06-07 13:34:51 -04:00
• 88be823d57 [AMD] Update compatible packaging version (#19309) pramenku 2025-06-07 18:25:09 +05:30
• 4e4f63ad45 [Nit][Benchmark]Fix example in benchmark_serving_structured_output.py (#19311) Lifans 2025-06-07 03:25:38 -07:00
• d2f0e7e615 [CI/Build] Improve Llama GGUF test robustness (#19287) Isotr0py 2025-06-07 17:23:28 +08:00
• 122cdca5f6 [Misc] refactor context extension (#19246) Reidandreidliu41 2025-06-07 13:13:21 +08:00
• cf02f9b283 Add FlexAttention to V1 (#16078) Driss Guessous 2025-06-07 00:58:55 -04:00
• c4296b1a27 [CI][PowerPC] Use a more appropriate way to select testcase in tests/models/language/pooling/test_embedding.py (#19253) Aaruni Aggarwal 2025-06-07 09:22:52 +05:30
• 66c508b137 [TPU][Test] Add script to run benchmark on TPU for buildkite (#19039) QiliangCui 2025-06-06 20:10:24 -07:00
• 84166fee97 [Kernel] Integrate CUTLASS MoE kernel with PPLX (#18762) ElizaWszolaandTyler Michael Smith 2025-06-07 03:26:11 +02:00
• 6e0cd10f72 [Easy][Test] Simplify test_function_tool_use with multiple parametrizes (#19269) Lu Fang 2025-06-07 09:19:09 +08:00
• e010688f50 [Build][ROCm] Update Dockerfile.rocm (#19296) Alexei-V-Ivanov-AMD 2025-06-06 18:35:16 -05:00
• 441b65d8c7 [Misc][Tools][Benchmark] Fix and improve auto tune script (#19163) Chenyaaang 2025-06-06 16:31:19 -07:00
• 46ecc57973 [BugFix] Fix tpu_model_runner block_id concatenation (#19228) Nick Hill 2025-06-06 16:28:17 -07:00
• b6a3a9f76d [Core] Fix abrupt request abort (#18485) Nicolò LucchesiandNick Hill 2025-06-07 01:27:59 +02:00
• ca27f0f9c1 [Bugfix][Core] Update cancellation logic in generate() to handle Generator exits (#19225) Adolfo VictoriaandAdolfo Victoria 2025-06-06 13:17:54 -07:00
• aad30bd306 [BugFix] Fix MultiConnector test after HMA changes (#19291) Nick Hill 2025-06-06 13:16:24 -07:00
• 94ecee6282 Fixed ppc build when it runs on non-RHEL based linux distros (#18422) NishidhaandMd. Shafi Hussain 2025-06-07 00:24:26 +05:30
• 8267f9916f improve logits bias (#19041) Yu Guo 2025-06-06 04:59:25 -07:00
• 7353492a47 [Core] Raise when non-multi-instance DP clients target a DP rank (#19227) jmswen 2025-06-06 04:03:01 -07:00
• 7661e92ef8 [Model] Optimize nemotron_h implementation (#19249) Jee Jee Li 2025-06-06 18:05:14 +08:00
• f168b85725 Unit Test for run_dp_sharded_vision_model (#19103) Siqi YanandSiqi Yan 2025-06-06 01:24:02 -07:00
• da511d54d8 Fix CompilationConfig repr (#19091) Richard Zou 2025-06-06 04:23:35 -04:00
• 65c69444b1 [Docs] Improve V1 KVConnector interface documentation (#19172) Nick Hill 2025-06-06 01:22:45 -07:00
• 94870359cd [Quantization] Bump compressed-tensors version; update NVFP4A16 test model (#19224) Dipika Sikka 2025-06-06 04:21:54 -04:00
• 0d49483ea9 [TPU] fix kv cache dtype in model runner (#19244) Chengji Yao 2025-06-06 01:20:16 -07:00
• 90b78ec5f9 [v1][P/D] Fix a edge case in kv cache schedule (#19182) Jinghui Zhangandjinghui 2025-06-05 23:32:55 -07:00
• 91a2ef98ea [Chore] update CODEOWNERS (#19247) Aaron Pham 2025-06-06 02:09:43 -04:00
• 3da2313d78 Support allowed_token_ids in ChatCompletionRequest (#19143) Xu Song 2025-06-06 13:06:48 +08:00
• b61dc5f972 [TPU] update torch_xla pin (#19231) Chengji Yao 2025-06-05 21:27:38 -07:00
• f8a1a2d108 [v1] Hybrid Memory Allocator (#17996) Chen Zhang 2025-06-06 11:47:09 +08:00
• 3465b87ef8 [Bugfix] Fix EAGLE vocab embedding construction for Llama 70B (#19033) Benjamin Chislett 2025-06-05 22:10:08 -04:00
• c8134bea15 Fix AOPerModuleConfig name changes (#18869) Jerry Zhang 2025-06-05 21:51:32 -04:00
• cb6d572e85 [Model] NemotronH support (#18863) Luis VegaandLuis Vega 2025-06-05 14:29:28 -07:00
• 87360308b7 [V1] Use FlashInfer by default on Blackwell GPUs (#19118) Michael Goin 2025-06-05 15:40:39 -04:00
• aa49f14832 [Quantization] Skip Fp4 Test for compressed-tensors (#19217) Dipika Sikka 2025-06-05 14:21:53 -04:00
• 9ef9173cfa [P/D][NixlConnector] Enable FlashInfer backend (#19090) Nicolò Lucchesi 2025-06-05 19:10:15 +02:00
• 85e2b7bb13 [MISC][Bugfix] Use less CPU when message queue has been empty for some time (#16226) Povilas Kanapickas 2025-06-05 19:53:08 +03:00
• 61059bee40 [Hardware][NVIDIA] FP4 MoE kernel optimization (#19110) Chiyue WeiandChiyue Wei 2025-06-05 09:48:26 -07:00
• ec89524f50 Add H20-3e fused MoE kernel tuning configs for DeepSeek-R1/V3 (#19205) Xu Wenqing 2025-06-06 00:38:54 +08:00
• f20f9f063b [mistral_common] Add v11 tokenizer (#19193) Patrick von Platen 2025-06-05 17:27:41 +02:00
• 9bc8bb07cf [Bugfix] properly catch PIL-related errors for vision models when incorrect data urls are provided (#19202) Guillaume Calmettes 2025-06-05 14:59:28 +02:00
• 1aeb925f34 [Frontend] improve vllm run-batch --help display (#19187) Reidandreidliu41 2025-06-05 19:16:25 +08:00
• 188a4590d8 [Misc] Do not override NCCL_CUMEM_ENABLE if set explicitly (#19105) 22quinn 2025-06-05 04:14:32 -07:00
• 18093084be [Misc] Remove unnecessary fallback to prefill-decode attention (#19138) vllmellm 2025-06-05 16:08:26 +08:00
• da40380214 [Build] Annotate wheel and container path for release workflow (#19162) Simon Moandgemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> 2025-06-04 23:24:56 -07:00
• 8fc57501d3 [Bugfix]: Fix the incompatibility issue with stream when Thinking is disabled (#19135) Chauncey 2025-06-05 14:24:24 +08:00
• af7fc84fd2 [BugFix][Minor] Fix full cuda graph bug when max_num_seqs < 512 (#19171) Woosuk Kwon 2025-06-04 22:41:25 -07:00
• 0678b52251 Handle non-serializable objects when dumping benchmark results (#19114) Huy Do 2025-06-04 22:40:04 -07:00
• 25b918eee6 [Torch Nightly]add missing dependency (#18770) Yang Wang 2025-06-04 21:56:12 -07:00
• a408820f2f [Bugfix] Fix port handling in make_zmq_path (#19117) Michael Goin 2025-06-04 23:00:59 -04:00