Commit Graph

Select branches

Hide Pull Requests

main

patches

sm86

Mono Color

• 7c3a0741c6 [Bugfix] Fix PixtralHFImagePixelInputs dynamic shape check (#22827) Isotr0py 2025-08-14 17:35:43 +08:00
• 00e3f9da46 vLLM Benchmark suite improvement (#22119) Louie TsaiandLi, Jiang <bigpyj64@gmail.com> 2025-08-14 00:12:17 -07:00
• a353bd083d [CI] remove flaky v0 test (#22864) Robert ShawandRobert Shaw 2025-08-14 00:41:51 -04:00
• 1d20c34717 [CI] Fix tests/distributed/test_ca_buffer_sharing.py (#22849) 3 people 2025-08-14 05:09:30 +02:00
• b6af24fba7 [CI][Entrypoints]: add filter to generation to filter out invalid tool calls (#22826) Will Eaton 2025-08-13 23:09:07 -04:00
• 0ca2393b47 [CI/Build] Increase pooling tolerance to pass CI (#22844) Cyrus Leungandmgoin 2025-08-14 06:52:48 +08:00
• 31a500c86f [Core] [N-gram SD Optimization][1/n] Propose tokens with a single KMP (#22437) Jialin Ouyang 2025-08-13 14:44:06 -07:00
• 4e8614e88b Move checklist in PR template (#22852) Luka Govedič 2025-08-13 17:38:35 -04:00
• c6cd5ca3d3 [ROCm][Bugfix] Fix compilation error in topk softmax fused kernel (#22819) kliuae 2025-08-14 04:45:03 +08:00
• df0e0f023e [CI/Build] Skip gpt_big model test because of broken HF model (#22848) Isotr0py 2025-08-14 04:36:28 +08:00
• b4b78d6317 [CI/Build] Fix param mismatch in test_eagle_correctness (#22847) Cyrus Leung 2025-08-14 01:55:25 +08:00
• 12817a8ac7 [CI] Fix tests/v1/e2e/test_kv_sharing_fast_prefill.py import on test (#22815) Nicolò Lucchesi 2025-08-13 19:35:50 +02:00
• c9232d41f4 [CI/Build] Update VLM common tests (#22841) Cyrus Leung 2025-08-14 01:03:05 +08:00
• 9bd9294f0e [Bugfix] Fix MiniCPMV Image input inference failed (#22813) 3 people 2025-08-14 00:41:41 +08:00
• da2705198f [Misc] clear and separate error messages for input too long and input + max-tokens too long (#22803) Roger Wang 2025-08-13 07:22:56 -07:00
• 19b927e52d [Core] Use individual MM items in P0/P1 cache and model runner (#22570) Cyrus Leung 2025-08-13 22:18:07 +08:00
• 20d65aa755 [Frontend] Multithreaded async multimodal load_bytes (#22710) milesialandAlexandre Milesi 2025-08-13 06:09:26 -07:00
• b159c0a67a Fix GGUF loader for Qwen3 MoE. (#22785) Gh0u1L5 2025-08-13 21:08:23 +08:00
• 6772bb0f7d Remove unnecessary CUDA sync of qwen image and video preprocess (#22792) Yuanyuan ChenandCyrus Leung 2025-08-13 21:07:28 +08:00
• fceafaf582 [Bugfix][mamba] Fix type annotation of Mamba2Metadata (#22787) Chen Zhang 2025-08-13 06:07:09 -07:00
• 6b794c756c [Nixl][CI] Fix tests (#22806) Nicolò Lucchesi 2025-08-13 15:03:53 +02:00
• 98deac3879 [FEATURE] support custom vllm tuned config path for fused moe triton kernels (#22791) Chi Zhang 2025-08-13 20:27:25 +08:00
• 653124bd46 [Frontend] Add chunked processing to handle long inputs in embedding models (#22280) 4 people 2025-08-13 19:14:24 +08:00
• 0b1bdac6af [Platform] Custom ops support for FusedMoe (#22509) wangxiyuan 2025-08-13 19:12:00 +08:00
• d94e3026de [V1] Add tree drafting tests for eagle spec decoding (#22705) Giancarlo Delfin 2025-08-13 04:11:28 -07:00
• 3f52738dce [Doc] Add max_lora_rank configuration guide (#22782) 633WHU 2025-08-13 19:10:07 +08:00
• a01e0018b5 [Bugfix] Fix Nemotron VL image processing (#22739) Duc-Viet Hoangandducviet00-h2 2025-08-13 17:11:36 +07:00
• 9e7e5baaa8 [Model] Add missing prefix to glm4_1v (#22716) Yuxuan Zhang 2025-08-13 16:23:33 +08:00
• d16aa3dae4 [Model] Add option to run Step3VisionEncoder in DP (#22697) zzh142857 2025-08-13 03:09:13 -04:00
• 6807af8f46 [gpt-oss] upgrade gpt-oss to v0.0.3 and add version check (#22768) Chen Zhang 2025-08-12 21:37:26 -07:00
• 4c558cf62e [Perf] Support topk softmax fused kernel for broader num_experts (#22211) shixiancandShixian Cui 2025-08-12 21:34:47 -07:00
• 77a6bf07ae [Bug] Fix Unexpected Keyword Argument 'w1_bias' (#22757) Wentao Ye 2025-08-13 00:31:47 -04:00
• 4082338a25 Remove unneeded ROCm platform import when using CUDA (#22765) Michael Goin 2025-08-13 00:26:38 -04:00
• c6b928798e Force TRTLLM attention for gpt-oss on SM100 (#22678) Michael Goin 2025-08-13 00:22:16 -04:00
• b1361c7273 [Bugfix] Fix default enable for CUTLASS MLA on SM100 (#22738) Michael Goin 2025-08-13 00:22:05 -04:00
• 4f0f844b16 Fix cuda illegal mem access with Llama4 TP8 + rms_norm custom op (#22701) Po-Han Huang (NVIDIA) 2025-08-13 12:21:50 +08:00
• c5830381af [V0 Deprecation] Remove args for multi-step scheduling (#22779) Woosuk Kwon 2025-08-12 20:38:18 -07:00
• d31f97cf57 [Misc] Remove tests/multi_step/__init__.py (#22778) Woosuk Kwon 2025-08-12 20:21:18 -07:00
• 71683ca6f6 [V0 Deprecation] Remove multi-step scheduling (#22138) Woosuk Kwon 2025-08-12 20:18:39 -07:00
• e18859298d Add hardware plugins to installation doc (#22732) Michael GoinandHarry Mellor 2025-08-12 20:14:46 -04:00
• fde0b611a3 [Model] Decouple glm4v (#22751) Jee Jee Li 2025-08-13 08:13:17 +08:00
• d0a6301588 Fix Transformers backend tensor parallel for multimodal models (#22673) Harry Mellor 2025-08-13 01:12:30 +01:00
• 45c3936e94 [Docs] Hide the navigation and toc sidebars on home page (#22749) Harry Mellor 2025-08-13 01:12:26 +01:00
• ba81acbdc1 [Bugfix] Bump DeepGEMM Version to Fix SMXX Layout Issues (#22606) Frank Wang 2025-08-12 15:43:06 -07:00
• 53c730286c [Misc] parametrize 'dtype' in test_flash_mla (#22641) RUTHLESS-BOTandgemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> 2025-08-13 04:31:48 +08:00
• 6534d2fc97 Fix torch version check for SM100 mxfp4 (#22535) zifeitongandmgoin 2025-08-12 12:54:42 -07:00
• 422f22e012 [CI][Nixl] Check kv cache layout during handshake (#22745) Nicolò Lucchesi 2025-08-12 21:53:52 +02:00
• 6bd8ebf026 [Kernel][AMD] Avoid D2H copy and cumsum kernel (#22683) 3 people 2025-08-12 12:53:36 -07:00
• dab4f9f764 [Chore] Update CODEOWNERS to include @yewentao256 for CUDA kernels, attention backends, quantization, and related tests (#22741) Wentao Ye 2025-08-12 12:50:31 -04:00
• c42fe0b63a Add more test scenario for tensor schema (#22733) TeeKen Lau 2025-08-13 02:34:41 +10:00
• 5a4b4b3729 Add: SupportsEagle3 interface for explicit EAGLE3 support (#22642) Rahul Tuli 2025-08-12 21:54:52 +05:30
• e5d3d63c42 [Benchmark] Fix terminal colors in benchmark_serving_multi_turn (python 3.12) (#22730) Daniel Serebrenik 2025-08-12 17:41:37 +03:00
• 3d9d40efde [Bugfix][CI] Fix test_remote_decode_lifecycle.py::test_short_prompt_lifecycle (#22727) Nicolò Lucchesi 2025-08-12 16:30:17 +02:00
• 67c153b88a Fix Llama4 FlashInfer FP4 MoE issues (#22511) Po-Han Huang (NVIDIA) 2025-08-12 20:50:59 +08:00
• f7ad6a1eb3 [CI Failure] fix tests/entrypoints/openai/test_skip_tokenizer.py (#22708) wang.yuqi 2025-08-12 20:42:58 +08:00
• 80bb1e8afe Officially support SmolLM3 using the Transformers backend (#22665) Harry Mellor 2025-08-12 13:38:48 +01:00
• d030b01548 [BugFix][Nixl][PD] Fix heterogenous TP (#22663) Nicolò LucchesiandNick Hill 2025-08-12 14:37:30 +02:00
• 767e63b860 [Docs] Improve docs navigation (#22720) Harry Mellor 2025-08-12 12:25:55 +01:00
• 007dd90859 [gpt-oss] Enable gpt-oss on ampere (#22714) Yongye Zhu 2025-08-12 06:21:44 -04:00
• b8a9d0e429 [Misc] remove GH discussions link (#22722) Jee Jee Li 2025-08-12 18:15:33 +08:00
• 50f2aae1b4 [LMCache][Example] Align the PYTHONHASHSEED for prefillers and decoders for KV chunks hashing (#21161) zejunchen-zejun 2025-08-12 17:05:14 +08:00
• 46ae7f6666 [Bugfix] Mamba2 SSD varlen bug fix initstates decay, improve test, assert chunk pwr 2 (#21783) RishiAstra 2025-08-12 05:04:37 -04:00
• 1ece7f30ba Fix: AWQ Marlin get_quant_method does not recognize "modules_to_not_convert" (#21888) 3 people 2025-08-12 17:03:53 +08:00
• bc8372efc3 [Bugfix] Fix erroneous randomly generated cases in bad word testing (#22170) phantomlei 2025-08-12 17:03:22 +08:00
• 8d17fa633e [V0] Correct CUDA Graph capture for encoder-decoder models (#22630) Sugar-zsg 2025-08-12 17:01:08 +08:00
• 9f909b8996 [New Model] Support Command-A-Vision (#22660) dongluw 2025-08-12 04:39:54 -04:00
• 59f3b93636 [DOC] update v1_guide with INTEL HW (#22679) Chendi.Xue 2025-08-12 03:22:49 -05:00
• 78077d5417 Move SchedulerConfig from config/__init__.py to config/scheduler.py (#22626) Harry Mellor 2025-08-12 08:23:49 +01:00
• 6d729c43fb [Bugfix] Fix ModernBert load & Enable sliding window attention for bidirectional attention. (#22637) wang.yuqiandMax de Bayser 2025-08-12 15:23:17 +08:00
• 2f4657952b [doc] Update x86 CPU-inference installation doc to reflect optionality of AVX512f (#22707) 3 people 2025-08-12 12:51:08 +05:30
• 3a7e3bbdd2 [Doc] Added unmentioned required option "method" in the usage of EAGLE-3 based models (#21737) Hongsheng LiuandDilute-l 2025-08-12 15:14:51 +08:00
• 4fbd8bb597 Fix passing SpeculativeConfig from the CLI (#22652) Harry Mellor 2025-08-12 06:13:32 +01:00
• ad344ef552 [gpt-oss] Small bug fixes for frontend (#22512) Chen Zhang 2025-08-11 22:04:38 -07:00
• bbaf9e9cb1 [gpt-oss] Fix mxfp4 support (#22700) Chen Zhang 2025-08-11 21:22:26 -07:00
• 4678503476 Migrate MiniCPMVImageInputs to TensorSchema (#21939) Benji Beck 2025-08-11 20:43:37 -07:00
• 93d0652433 [CI] Increase timeout for test_completion_with_image_embeds (#22670) Michael Goin 2025-08-11 23:31:36 -04:00
• ea1292ad3e [CI Failure] Use float32 for tests/entrypoints/openai/test_audio.py (#22686) Michael Goin 2025-08-11 23:20:42 -04:00
• dc5e4a653c Upgrade FlashInfer to v0.2.11 (#22613) Po-Han Huang (NVIDIA)andmgoin 2025-08-12 10:58:41 +08:00
• 839ab00349 Re-enable Xet on TPU tests now that hf_xet has been updated (#22666) Harry Mellor 2025-08-12 03:54:40 +01:00
• 9b94d6ec8f Enable 4bit bnb prequant MOE (#21548) Andy ChenandJee Jee Li 2025-08-11 19:02:14 -07:00
• 1891a265d3 [gpt-oss] Add test for response API + harmony (but skipped) (#22554) Chen Zhang 2025-08-11 17:47:24 -07:00
• 95a935fc48 [gpt-oss] Support streaming in response API (#22431) Chen Zhang 2025-08-11 17:46:59 -07:00
• 458e74eb90 Support more parallel styles in Transformers backend TP (#22651) Harry Mellor 2025-08-11 18:42:48 +01:00
• 65abe111a3 [CI] Skip Tree Attn Test in test_max_len.py to unblock CI (#22664) TJian 2025-08-11 10:36:05 -07:00
• 807d21b80d [BugFix] [Spec Decode] Remove LlamaForCausalLMEagle3 to fix CI (#22611) 22quinn 2025-08-11 10:31:36 -07:00
• c90fb03df5 [CI/Build] Skip Mllama HF runner tests with Transformers v4.55.0 (#22659) Isotr0py 2025-08-12 01:00:58 +08:00
• 84cf78acee [Model] Pooling models default to using chunked prefill & prefix caching if supported. (#20930) wang.yuqi 2025-08-12 00:41:37 +08:00
• 16fb668b61 fix: NIXL connector transfers partial block to pass full multi-modal context (#21074) GuanLuo 2025-08-11 09:40:55 -07:00
• f7dcce7a4a [Feature] Add VLLM_USE_DEEP_GEMM_E8M0 Env to Control E8M0 Scale (#21968) Wentao Ye 2025-08-11 12:39:08 -04:00
• 8e13d9fe6d [Misc] Further clean up some redundant config definitions (#22649) Isotr0py 2025-08-12 00:22:25 +08:00
• 3fa5b25845 Document aarch64 CPU support works (#22646) Eric Curtin 2025-08-11 15:22:45 +01:00
• 14a5d903ab [Model] NemotronH Support (#22349) danielafrimi 2025-08-11 14:09:24 +03:00
• 951b038298 [Misc] Move jsontree to utils (#22622) Cyrus Leung 2025-08-11 18:49:32 +08:00
• ebf7605b0d [Misc] Move tensor schema tests (#22612) Cyrus Leung 2025-08-11 15:15:27 +08:00
• bc1d02ac85 [Docs] Add comprehensive CLI reference for all large vllm subcommands (#22601) Harry Mellor 2025-08-11 08:13:33 +01:00
• 1e55dfa7e5 [BUGFIX] KeyError 'layers.14.mlp.gate.g_idx' for Qwen3-MoE with GPTQ on ROCm (#22017) JartX 2025-08-11 09:13:30 +02:00
• 384a052971 [Misc] benchmark_moe supports expert parallel (#22251) Jee Jee Li 2025-08-11 15:13:27 +08:00
• 39052dbca8 Support token_type_ids in V1 with less code changes (#21985) Maximilien de Bayser 2025-08-11 02:54:59 -03:00
• 9c97a1c349 [ROCm][AITER] Support AITER Rope ops in RotaryEmbedding Module. (#22521) vllmellm 2025-08-11 13:52:34 +08:00
• f919d4cb8f [BugFix] Fix logits repetition penalty cuda check (#22592) Eugene Cheah 2025-08-10 22:52:31 -07:00