Commit Graph

Select branches

Hide Pull Requests

main

patches

sm86

Mono Color

• 4b0da7b60e Enable hybrid attention models for Transformers backend (#18494) Harry Mellor 2025-05-23 04:12:08 +02:00
• c6b636f9fb [V1][Spec Decoding] Use model_loader.get_model() to load models (#18273) Mark McLoughlin 2025-05-23 03:05:44 +01:00
• 04eb88dc80 Re-submit: Fix: Proper RGBA -> RGB conversion for PIL images. (#18569) Chenheli Hua 2025-05-22 18:59:18 -07:00
• 46791e1b4b [AMD] [P/D] Compute num gpus for ROCm correctly in run_accuracy_test.sh (#18568) rasmith 2025-05-22 20:45:35 -05:00
• c32e249a23 [Frontend] [Core] Add Tensorizer support for V1, LoRA adapter serialization and deserialization (#17926) Sanger Steel 2025-05-22 21:44:18 -04:00
• c91fe7b1b9 [Frontend][Bug Fix] Update llama4 pythonic jinja template and llama4_pythonic parser (#17917) Kai Wu 2025-05-22 16:44:08 -07:00
• a04720bc36 [V1][Spec Decode][Bugfix] Load quantize weights for EAGLE (#18290) Ekagra Ranjan 2025-05-22 18:17:33 -04:00
• 7b9d832c80 [Tool] Add NIXL installation script (#18172) lkchen 2025-05-22 14:33:16 -07:00
• 6e588da0f4 [Build/CI] Fix CUDA 11.8 build (#17679) Tyler Michael SmithandLucas Wilkinson 2025-05-22 15:13:54 -04:00
• f8d2cc5f55 [Compile][Platform] Make PiecewiseBackend pluggable and extendable (#18076) Mengqing Caoandyoukaichao 2025-05-23 03:11:53 +08:00
• 721fb9b181 [Platform] Move platform check to right place (#18470) wangxiyuan 2025-05-23 03:11:28 +08:00
• 1f3a1200e4 [Bugfix] make test_openai_schema.py pass (#18224) David XiaandHarry Mellor 2025-05-22 14:34:06 -04:00
• 54631f8262 [Misc] Call ndarray.tobytes() directly instead of ndarray.data.tobytes() (#18347) Lukas Geiger 2025-05-22 17:00:13 +01:00
• cb506ecb5a [Misc] improve Automatic Prefix Caching example (#18554) Reidandreidliu41 2025-05-22 22:50:46 +08:00
• 93f71673ce [BugFix][CPU] Fix x86 SHM distributed module initialization (#18536) Li, Jiang 2025-05-22 22:35:00 +08:00
• 3f505233fd [Doc] Add stream flag for chat completion example (#18524) Calvin Chen 2025-05-22 22:07:10 +08:00
• 4e04eceb58 [Bugfix] Use random hidden states in dummy sampler run (#18543) Bowen Wang 2025-05-22 06:48:56 -07:00
• 71075029f2 [Doc] Support --stream arg in openai_completion_client.py script (#18388) CYJiang 2025-05-22 21:20:17 +08:00
• ca86a7cf6e [CI/Build] Update bamba test model location (#18544) Harry Mellor 2025-05-22 15:01:07 +02:00
• a35a494745 [Bugfix] Add kwargs to RequestOutput __init__ to be forward compatible (#18513) lkchen 2025-05-22 05:24:43 -07:00
• f6037d1907 [Bugfix] Fix MRoPE Errors in the Qwen-VL Model When Processing Pure Text (#18526) 4 people 2025-05-22 20:22:53 +08:00
• fa72f9a812 Order sequence ids + config update to support specifying custom quantization layers (#18279) 7 people 2025-05-22 02:20:36 -07:00
• ebed81fbf5 Update default neuron config for speculation (#18274) 3 people 2025-05-22 02:18:55 -07:00
• e2d7d31244 [Neuron] Update Dockerfile.neuron to use latest neuron release (2.23) (#18512) Satyajith Chilappagari 2025-05-22 02:17:34 -07:00
• 23b67b37b2 [Doc] Fix invalid JSON in example args (#18527) Cyrus Leung 2025-05-22 15:11:46 +08:00
• db5a29ba19 [Bugfix] Fix LoRA test (#18518) Jee Jee Li 2025-05-22 12:48:53 +08:00
• 51797775c3 [Bugfix][Model] Make Olmo2Model weight loading return loaded weights (#18504) Shane A 2025-05-21 21:17:03 -07:00
• cf5984b2fe [BugFix][DP] Send DP wave completion only from dp_rank==0 (#18502) Nick Hillandkourosh hakhamaneshi 2025-05-21 20:25:25 -07:00
• d022115cc6 [Bugfix] Inconsistent token calculation compared to HF in llava family (#18479) youngrok cha 2025-05-22 12:21:47 +09:00
• acb54ca8e1 Intialize io_thread_pool attribute in the beginning. (#18331) Rabi Mishra 2025-05-22 08:51:14 +05:30
• 6e0fd34d3c [CI] Fix race condition with StatelessProcessGroup.barrier (#18506) Russell Bryant 2025-05-21 23:19:13 -04:00
• 176d62e4ea [MISC] update project urls in pyproject.toml (#18519) Ning Xie 2025-05-22 11:17:34 +08:00
• 20bd6f4d2e [FalconH1] Fix output dtype in RMSNorm fallback path for Falcon-H1 (e.g. 0.5B) (#18500) 4 people 2025-05-22 06:23:59 +04:00
• 1f079540db [Bugfix] Consistent ascii handling in tool parsers (#17704) Sebastian Schoennenbeck 2025-05-21 22:41:23 +02:00
• 94d8ec8d2b [FEAT][ROCm] Upgrade AITER MLA v1 backend (#18338) vllmellmandLuka Govedič 2025-05-22 01:34:28 +08:00
• bb0a311213 Revert "[v1] Support multiple KV cache groups in GPU model runner (#17945) (#18459) Mark McLoughlin 2025-05-21 18:25:23 +01:00
• dd5fa7e04f [ROCm][Kernel][V1] Enable AMD Radeon GPU Custom Paged Attention on v1 (#17004) Hosang 2025-05-21 11:35:00 -04:00
• 2b16104557 [Misc] Update deprecation message for --enable-reasoning (#18404) Hyogeun Oh (오효근) 2025-05-21 23:33:11 +09:00
• 371376f996 [Build] fix Dockerfile shell (#18402) Kebe 2025-05-21 22:32:06 +08:00
• c6c10ca920 [Bugfix] Reduce moe_sum test size to avoid OOM (#18484) bnellnm 2025-05-21 09:46:39 -04:00
• c154d89306 [Doc] fix arg docstring in linear layers (#18410) GiantCroc 2025-05-21 21:45:57 +08:00
• eca18691d2 [MODEL] FalconH1 (#18406) 4 people 2025-05-21 15:59:06 +04:00
• 61acfc45bc [Bugfix][Failing Test] Fix test_events.py (#18460) Rabi Mishra 2025-05-21 17:27:28 +05:30
• 107f5fc4cb [Misc] refactor disaggregated-prefill-v1 example (#18474) Reidandreidliu41 2025-05-21 19:10:14 +08:00
• 907f935de9 [V1] Fix general plugins not loaded in engine for multiproc (#18326) Yong Hoon Shin 2025-05-21 01:21:49 -07:00
• 5d7f545204 [Frontend] deprecate --device arg (#18399) Kebe 2025-05-21 16:21:17 +08:00
• cd8dfc6dfc [Misc] MultiConnector._connectors type (#18423) Nicolò Lucchesi 2025-05-21 07:48:43 +02:00
• d06dd72ba9 [Bugfix][Failing Test] Fix nixl connector test when promt size < block size (#18429) wwl2755 2025-05-21 00:41:44 -05:00
• ad0012a0ac Revert "[Bugfix] Fix MRoPE Errors in the Qwen-VL Model When Processing Pure Text (#18407)" (#18456) Cyrus Leung 2025-05-21 13:39:22 +08:00
• 92247c522e [Bug] Fix moe_sum signature (#18440) bnellnm 2025-05-21 01:37:08 -04:00
• 0c15c2e486 [Bugfix] config.head_dim is now explicitly set to None (#18432) Gregory Shtrasberg 2025-05-21 00:04:33 -04:00
• 3b17ea26e4 [TPU] Re-enable the Pallas MoE kernel (#18025) Michael Goin 2025-05-20 22:52:27 -04:00
• 23baa2180b fix:Build torch wheel inline rather than picking from nightly (#18351) Dilip Gowda Bhagavan 2025-05-21 03:52:24 +05:30
• 980a172474 [Kernel] update comment for KV shape in unified triton attn (#18099) Percy 2025-05-20 13:19:34 -05:00
• e1f5a71ed7 [Model] use AutoWeightsLoader for bloom (#18300) Calvin Chen 2025-05-21 00:40:05 +08:00
• f4a8a37465 [Minor] Rename quantization nvfp4 to modelopt_fp4 (#18356) Michael Goin 2025-05-20 12:08:37 -04:00
• 8f55962a7f [Misc] refactor prompt embedding examples (#18405) Reidandreidliu41 2025-05-20 23:26:12 +08:00
• be48360c1f [Bugfix] Fix MRoPE Errors in the Qwen-VL Model When Processing Pure Text (#18407) 燃and松灵 2025-05-20 21:59:48 +08:00
• 86847700d7 [CI] Add mteb testing to test the accuracy of the embedding model (#17175) wang.yuqi 2025-05-20 21:51:12 +08:00
• d6c86d09ae Update cpu.txt (#18398) 汪志鹏 2025-05-20 18:53:23 +08:00
• 6b35cb10a0 [Misc] Add LoRA code owner (#18387) Jee Jee Li 2025-05-20 18:27:30 +08:00
• 1b1e8e05ff [doc] update env variable export (#18391) Reidandreidliu41 2025-05-20 16:53:27 +08:00
• bca55b556f [Bugfix] fix adding bias twice in ipex GPTQ quantization (#18363) Random Fly 2025-05-20 15:54:33 +08:00
• d981396778 [release] Change dockerhub username for TPU release (#18389) Kevin H. Luu 2025-05-19 23:49:23 -07:00
• 9609327fa4 [Core] [Bugfix]: tensor parallel with prompt embeds (#18171) Nan QinandAndrew Sansom 2025-05-19 22:21:27 -05:00
• f07a673eb2 [Misc] Allow AutoWeightsLoader to skip loading weights with specific substr in name (#18358) Isotr0py 2025-05-20 11:20:12 +08:00
• d565e0976f [neuron] fix authorization issue (#18364) Liangfu Chen 2025-05-19 16:30:32 -07:00
• 258bf621d5 fix CUDA_check redefinition in #17918 (#18287) Lucia FangandLucia Fang 2025-05-19 13:42:35 -07:00
• dc1440cf9f Neuron up mistral (#18222) Satyajith Chilappagari 2025-05-19 09:54:47 -07:00
• 8171221834 [Misc] Fix typo (#18330) Gong Shufan 2025-05-20 00:51:01 +08:00
• 7937c2fd52 Add files via uploadAdd fused MoE kernel tuning configs (fp8_w8a8) for DeepSeek V3/R1 on a single-node 8x NVIDIA H20 96GB setup (#18337) sunyicode0012 2025-05-20 00:49:57 +08:00
• e2ee1e8e9e [Feature]Add support for models quantized with AutoRound (#17850) Wenhua Cheng 2025-05-20 00:38:53 +08:00
• 20d8ce81eb [Frontend] add --quick option for vllm chat/complete (#18297) Reidandreidliu41 2025-05-20 00:36:13 +08:00
• 84ab4feb7e [Doc] Fix typo (#18355) Elad Segal 2025-05-19 19:05:16 +03:00
• 6781af5608 [Quantization] Pool model support bitsandbytes (#18087) Jee Jee Li 2025-05-20 00:03:43 +08:00
• 1b15df2546 [BugFix] Fix handling of num_computed_tokens with connector (#18232) Nick HillandNicolò Lucchesi 2025-05-19 09:03:25 -07:00
• 43b5f61dce [Doc] Move input-related docs to Features (#18353) Cyrus Leung 2025-05-19 23:08:39 +08:00
• c5bb0ebdc6 [Doc] Fix prompt embedding examples (#18350) Li Wang 2025-05-19 21:48:16 +08:00
• d637b96099 [BugFix] [Vul] Add missing usedforsecurity=False in MD5 hashing to enable FIPS (#18319) Shaoyu Yangandcascade 2025-05-19 16:31:23 +08:00
• 275c5daeb0 fix: Add type specifications for CLI arguments in tensorizer options (#18314) CYJiang 2025-05-19 14:42:17 +08:00
• 47fda6d089 [Build] Supports CUDA 12.6 and 11.8 after Blackwell Update (#18316) Simon Mo 2025-05-18 23:19:33 -07:00
• 27d0952600 [Misc] extract parser.parse_args() (#18323) Reidandreidliu41 2025-05-19 12:06:26 +08:00
• 221cfc2fea Feature/vllm/input embedding completion api (#17590) 6 people 2025-05-18 22:18:05 -05:00
• 9da1095daf [Spec Decode][V0] Fix spec decode correctness test in V0 eagle/medusa (#18175) wwl2755 2025-05-18 21:49:46 -05:00
• d1211f8794 [Doc] Add doc to explain the usage of Qwen3 thinking (#18291) Robin 2025-05-19 07:04:07 +08:00
• b6a6e7a529 [Misc] add litellm integration (#18320) Reidandreidliu41 2025-05-18 23:32:30 +08:00
• 4fb349f66a Fix copy-paste error in phi4mm image processing (#18315) Lifu Huang 2025-05-18 07:00:12 -07:00
• 908733aca7 [Model] Use sigmoid for single-label classification (#18313) 22quinn 2025-05-18 07:00:09 -07:00
• 1a8f68bb90 [doc] update reasoning doc (#18306) Reidandreidliu41 2025-05-18 21:59:14 +08:00
• 9ab2c02ff8 Support sequence parallelism combined with pipeline parallelism (#18243) cascade 2025-05-17 15:47:25 -07:00
• 66e63e86ec [MISC] fix typo (#18305) Ning Xie 2025-05-18 01:52:09 +08:00
• 9214e60631 [Model] use AutoWeightsLoader for solar (#18113) rongfu.leng 2025-05-17 15:24:17 +08:00
• f880d42582 Fixed build on ppc64le due to openssl conflicts (#18262) Nishidha 2025-05-17 12:53:46 +05:30
• dcfe95234c Update Dockerfile to build for Blackwell (#18095) Michael Goin 2025-05-17 03:23:25 -04:00
• 48ac2bed5b [Hardware][TPU] Optionally import for TPU backend (#18269) 4 people 2025-05-17 00:23:12 -07:00
• 3e0d435027 [P/D][V1] Support dynamic loading of external KV connector implementations (#18142) David Ben-DavidandDavid Ben-David 2025-05-17 09:40:39 +03:00
• 4ee4826ede [BugFix] Correct max_model_len derivation from config.json for Mistral format (#17937) 3 people 2025-05-16 21:20:13 -07:00
• 60017dc841 [Misc] reformat the collect-env output (#18285) Reidandreidliu41 2025-05-17 10:46:18 +08:00
• 55f1a468d9 Move cli args docs to its own page (#18228) (#18264) Trevor Royer 2025-05-16 19:43:45 -07:00
• fd195b194e [V1][P/D] Local attention optimization for NIXL (#18170) Michael Goin 2025-05-16 21:16:33 -04:00