[{"data":1,"prerenderedAt":627},["ShallowReactive",2],{"repo-vllm-project-vllm-zh-CN":3,"i18n-vllm-project-vllm-zh-CN":624},{"mirrorPool":4,"detail":13,"relatedProjects":515},[5,9],{"key":6,"label":7,"base":8,"type":6},"github","mirror.github","https:\u002F\u002Fgithub.com\u002F",{"key":10,"label":11,"base":12,"type":10},"auto-proxy","mirror.autoProxy","https:\u002F\u002Fgithub.xxlab.tech\u002F",{"owner":14,"repo":15,"repoId":16,"repoUrl":17,"starCount":18,"language":19,"description":20,"languages":21,"screenshots":30,"homepage":31,"pushedAt":32,"ownerAvatar":33,"ownerName":34,"ownerBio":35,"lastSyncTime":36,"lastReleaseTime":37,"tags":38,"license":59,"category":60,"subgroup":61,"releases":62},"vllm-project","vllm",599547518,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm",94207,"Python","A high-throughput and memory-efficient inference and serving engine for LLMs",{"Python":22,"Rust":23,"Cuda":24,"C++":25,"Shell":26,"CMake":27,"Dockerfile":28,"HCL":29,"C":29,"Jinja":29},85.4,6.4,4,3,1,0.2,0.1,0,[],"https:\u002F\u002Fvllm.ai","2026-10-20","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F136984999?v=4","vLLM","","2026-10-20T10:23:42.503958+00:00","2026-10-18T06:50:48Z",[39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58],"amd","blackwell","cuda","deepseek","deepseek-v3","gpt","gpt-oss","inference","kimi","llama","llm","llm-serving","model-serving","moe","openai","pytorch","qwen","qwen3","tpu","transformer","Apache-2.0","development","LLM推理\u002F微调工具",[63,123,183,243,303,363,417,462],{"tag":64,"publishAt":37,"semVer":65,"prerelease":67,"assets":68},"v0.32.0",[29,66,29],32,false,[69,75,81,87,93,99,105,111,117],{"fileName":70,"sizeMb":71,"sha256":72,"downloadCount":73,"originUrl":74},"vllm-0.32.0+cpu-cp38-abi3-macosx_11_0_arm64.whl",28.54,"c28cf0fede58f02c4886d2794217ec3c7e4bbd22dfeb16b4b32586a8f230bb75",113,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.32.0\u002Fvllm-0.32.0%2Bcpu-cp38-abi3-macosx_11_0_arm64.whl",{"fileName":76,"sizeMb":77,"sha256":78,"downloadCount":79,"originUrl":80},"vllm-0.32.0+cpu-cp38-abi3-manylinux_2_39_aarch64.whl",67.6,"d61609855dd3c0a2c731b64f815b24b10216035f1807845df049d7efabbefdb1",23,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.32.0\u002Fvllm-0.32.0%2Bcpu-cp38-abi3-manylinux_2_39_aarch64.whl",{"fileName":82,"sizeMb":83,"sha256":84,"downloadCount":85,"originUrl":86},"vllm-0.32.0+cpu-cp38-abi3-manylinux_2_39_x86_64.whl",147.53,"e87a047cc5ab47eb14565dfcd060393876b33e913479982a56fca6fca73ed97a",133,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.32.0\u002Fvllm-0.32.0%2Bcpu-cp38-abi3-manylinux_2_39_x86_64.whl",{"fileName":88,"sizeMb":89,"sha256":90,"downloadCount":91,"originUrl":92},"vllm-0.32.0+cu129-cp38-abi3-manylinux_2_28_aarch64.whl",508.54,"258418994620a9bf904544d20f90e2cf6963788614b88d26d1b429f5d46a5194",42,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.32.0\u002Fvllm-0.32.0%2Bcu129-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":94,"sizeMb":95,"sha256":96,"downloadCount":97,"originUrl":98},"vllm-0.32.0+cu129-cp38-abi3-manylinux_2_28_x86_64.whl",533.51,"9918be806fb2808c406547ff4f9f43870ac6ff28dc4b179291fcf532145bb50b",1359,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.32.0\u002Fvllm-0.32.0%2Bcu129-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":100,"sizeMb":101,"sha256":102,"downloadCount":103,"originUrl":104},"vllm-0.32.0+xpu-cp38-abi3-manylinux_2_34_x86_64.whl",31.97,"650f5c2627057333dbeab044bfd97b189e57c44ca3bfdddb02724691cdf94fc0",21,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.32.0\u002Fvllm-0.32.0%2Bxpu-cp38-abi3-manylinux_2_34_x86_64.whl",{"fileName":106,"sizeMb":107,"sha256":108,"downloadCount":109,"originUrl":110},"vllm-0.32.0-cp38-abi3-manylinux_2_28_aarch64.whl",306.55,"d0620dbdf99528bcfebe551a8906347a229921fb1602f887bd1286a920da2d44",715,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.32.0\u002Fvllm-0.32.0-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":112,"sizeMb":113,"sha256":114,"downloadCount":115,"originUrl":116},"vllm-0.32.0-cp38-abi3-manylinux_2_28_x86_64.whl",311.39,"5e2ecdc5b0baf72921c288486417ca82778456a448e4fe60b62ee43c0dfc8611",788,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.32.0\u002Fvllm-0.32.0-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":118,"sizeMb":119,"sha256":120,"downloadCount":121,"originUrl":122},"vllm-0.32.0.tar.gz",42.49,"d2c957d5eee3d0985012151315ddc62cbe0cd368fc96a5fc3510438d9dddf2ad",47,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.32.0\u002Fvllm-0.32.0.tar.gz",{"tag":124,"publishAt":125,"semVer":126,"prerelease":67,"assets":128},"v0.31.0","2026-10-05T06:44:55Z",[29,127,29],31,[129,135,141,147,153,159,165,171,177],{"fileName":130,"sizeMb":131,"sha256":132,"downloadCount":133,"originUrl":134},"vllm-0.31.0+cpu-cp312-cp312-macosx_11_0_arm64.whl",28.13,"21378516a51ff15029f71168e827b94660e21b39484eb52277166f01ac7dd686",5552,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.31.0\u002Fvllm-0.31.0%2Bcpu-cp312-cp312-macosx_11_0_arm64.whl",{"fileName":136,"sizeMb":137,"sha256":138,"downloadCount":139,"originUrl":140},"vllm-0.31.0+cpu-cp38-abi3-manylinux_2_39_aarch64.whl",65.46,"8a9f86447a8ebcd1dadeae81b4b52587d8d967f67500bc7c00d2c6383c1df9cb",4302,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.31.0\u002Fvllm-0.31.0%2Bcpu-cp38-abi3-manylinux_2_39_aarch64.whl",{"fileName":142,"sizeMb":143,"sha256":144,"downloadCount":145,"originUrl":146},"vllm-0.31.0+cpu-cp38-abi3-manylinux_2_39_x86_64.whl",143.89,"d1eb6f3b6e548cfe372ecc4c85d0d2aa6986f183a2bfcfccdff93549b5463227",11233,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.31.0\u002Fvllm-0.31.0%2Bcpu-cp38-abi3-manylinux_2_39_x86_64.whl",{"fileName":148,"sizeMb":149,"sha256":150,"downloadCount":151,"originUrl":152},"vllm-0.31.0+cu129-cp38-abi3-manylinux_2_28_aarch64.whl",502.71,"0220b1d3cc8833e32436a1bfa573915c861ccac171ccf6e5b9cbf27c7edddebc",4344,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.31.0\u002Fvllm-0.31.0%2Bcu129-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":154,"sizeMb":155,"sha256":156,"downloadCount":157,"originUrl":158},"vllm-0.31.0+cu129-cp38-abi3-manylinux_2_28_x86_64.whl",527.72,"5176ab746f4e736c4a354a11d0620abf0a941de287cad937ac34039293ca6f5f",12977,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.31.0\u002Fvllm-0.31.0%2Bcu129-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":160,"sizeMb":161,"sha256":162,"downloadCount":163,"originUrl":164},"vllm-0.31.0+xpu-cp38-abi3-manylinux_2_34_x86_64.whl",31.64,"d3e9e825505caed8071cdb725588878604a63ae56e987fb8bf2083b0421da6b7",106,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.31.0\u002Fvllm-0.31.0%2Bxpu-cp38-abi3-manylinux_2_34_x86_64.whl",{"fileName":166,"sizeMb":167,"sha256":168,"downloadCount":169,"originUrl":170},"vllm-0.31.0-cp38-abi3-manylinux_2_28_aarch64.whl",300.78,"56edd19c798510b6da5ae4dd38b3bdf55ea7ae261b339a906df006dc3faf751d",13523,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.31.0\u002Fvllm-0.31.0-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":172,"sizeMb":173,"sha256":174,"downloadCount":175,"originUrl":176},"vllm-0.31.0-cp38-abi3-manylinux_2_28_x86_64.whl",305.71,"39132b433967f0962d03d8a5f633b9bee6b32708f566c7d5637b79f8d0ef3433",14612,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.31.0\u002Fvllm-0.31.0-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":178,"sizeMb":179,"sha256":180,"downloadCount":181,"originUrl":182},"vllm-0.31.0.tar.gz",41.41,"dde46f1efbe846e5bd0c5c4ec6ec71842c42254c7ff3d9bebb2a57028715d380",241,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.31.0\u002Fvllm-0.31.0.tar.gz",{"tag":184,"publishAt":185,"semVer":186,"prerelease":67,"assets":188},"v0.30.0","2026-09-22T05:20:54Z",[29,187,29],30,[189,195,201,207,213,219,225,231,237],{"fileName":190,"sizeMb":191,"sha256":192,"downloadCount":193,"originUrl":194},"vllm-0.30.0+cpu-cp312-cp312-macosx_11_0_arm64.whl",27.63,"fd9adfd566a8afa4ecbdf36b04bab10ebf43cd9ca18eca4b0a0bf2013b8de135",6723,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.30.0\u002Fvllm-0.30.0%2Bcpu-cp312-cp312-macosx_11_0_arm64.whl",{"fileName":196,"sizeMb":197,"sha256":198,"downloadCount":199,"originUrl":200},"vllm-0.30.0+cpu-cp38-abi3-manylinux_2_39_aarch64.whl",65.16,"0423f802029362ca31940262c580ed3f5af9d10bd123106a47f0fde6296f3c47",611,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.30.0\u002Fvllm-0.30.0%2Bcpu-cp38-abi3-manylinux_2_39_aarch64.whl",{"fileName":202,"sizeMb":203,"sha256":204,"downloadCount":205,"originUrl":206},"vllm-0.30.0+cpu-cp38-abi3-manylinux_2_39_x86_64.whl",140.6,"0ee75278b3626c5d0b7c310c6d62afae93e900f4eac339c91e333fde5108ed78",3930,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.30.0\u002Fvllm-0.30.0%2Bcpu-cp38-abi3-manylinux_2_39_x86_64.whl",{"fileName":208,"sizeMb":209,"sha256":210,"downloadCount":211,"originUrl":212},"vllm-0.30.0+cu129-cp38-abi3-manylinux_2_28_aarch64.whl",495.89,"fdb57ab5fa1c3ac4c94a6eff52579df32cc9aab5da9863880d03e7167e088e77",359,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.30.0\u002Fvllm-0.30.0%2Bcu129-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":214,"sizeMb":215,"sha256":216,"downloadCount":217,"originUrl":218},"vllm-0.30.0+cu129-cp38-abi3-manylinux_2_28_x86_64.whl",520.19,"e98cb69659bfcfc849cf11ce0781a7161d40b02b51a6c3636924a5909f2aabcc",24193,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.30.0\u002Fvllm-0.30.0%2Bcu129-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":220,"sizeMb":221,"sha256":222,"downloadCount":223,"originUrl":224},"vllm-0.30.0+xpu-cp38-abi3-manylinux_2_34_x86_64.whl",31.13,"b5c09abc6a91c7b4430b7f36cd606ffd29e95b73c5f48f026bd2ebd2261124d6",98,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.30.0\u002Fvllm-0.30.0%2Bxpu-cp38-abi3-manylinux_2_34_x86_64.whl",{"fileName":226,"sizeMb":227,"sha256":228,"downloadCount":229,"originUrl":230},"vllm-0.30.0-cp38-abi3-manylinux_2_28_aarch64.whl",295.62,"eb3e11bab695d085098579a6eda2d602419adec3826ebfbcce9a3ffa543eb62e",9971,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.30.0\u002Fvllm-0.30.0-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":232,"sizeMb":233,"sha256":234,"downloadCount":235,"originUrl":236},"vllm-0.30.0-cp38-abi3-manylinux_2_28_x86_64.whl",300.3,"ef52ee58c410ead0b8afb190838fa4cbcb52075596f67862a03859d984966ac4",12177,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.30.0\u002Fvllm-0.30.0-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":238,"sizeMb":239,"sha256":240,"downloadCount":241,"originUrl":242},"vllm-0.30.0.tar.gz",40.47,"5f8f4e890c042ffa1c3e103f81c35e2d96f60a0a175ac43a4adbae7006bef62b",283,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.30.0\u002Fvllm-0.30.0.tar.gz",{"tag":244,"publishAt":245,"semVer":246,"prerelease":67,"assets":248},"v0.29.0","2026-09-09T08:54:49Z",[29,247,29],29,[249,255,261,267,273,279,285,291,297],{"fileName":250,"sizeMb":251,"sha256":252,"downloadCount":253,"originUrl":254},"vllm-0.29.0+cpu-cp312-cp312-macosx_11_0_arm64.whl",26.66,"7133cb494664c502b07b114fe915847f0e71296d502f67f6fa76172ba46978df",3454,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.29.0\u002Fvllm-0.29.0%2Bcpu-cp312-cp312-macosx_11_0_arm64.whl",{"fileName":256,"sizeMb":257,"sha256":258,"downloadCount":259,"originUrl":260},"vllm-0.29.0+cpu-cp38-abi3-manylinux_2_34_aarch64.whl",65.14,"f14a4c18c0ba55f6932f41064f8fe345387e99e3ec76bb9c4a63b6069e876ca0",7536,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.29.0\u002Fvllm-0.29.0%2Bcpu-cp38-abi3-manylinux_2_34_aarch64.whl",{"fileName":262,"sizeMb":263,"sha256":264,"downloadCount":265,"originUrl":266},"vllm-0.29.0+cpu-cp38-abi3-manylinux_2_34_x86_64.whl",131.47,"4d22dac8259e7e24dbc615531c73cbd1a6d4b4214a3329ccddb7854f26c9dac8",15224,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.29.0\u002Fvllm-0.29.0%2Bcpu-cp38-abi3-manylinux_2_34_x86_64.whl",{"fileName":268,"sizeMb":269,"sha256":270,"downloadCount":271,"originUrl":272},"vllm-0.29.0+cu129-cp38-abi3-manylinux_2_28_aarch64.whl",497.53,"b4efb6f657a1cc2ea8f898d39127974c53bd72206f95a5879611ade5c986e61f",255,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.29.0\u002Fvllm-0.29.0%2Bcu129-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":274,"sizeMb":275,"sha256":276,"downloadCount":277,"originUrl":278},"vllm-0.29.0+cu129-cp38-abi3-manylinux_2_28_x86_64.whl",523.08,"22e8d8fec755986b3ad964004a1f8c65a55626ec948354bdbe95993e6b0289fe",10172,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.29.0\u002Fvllm-0.29.0%2Bcu129-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":280,"sizeMb":281,"sha256":282,"downloadCount":283,"originUrl":284},"vllm-0.29.0+xpu-cp38-abi3-manylinux_2_34_x86_64.whl",30.13,"aed7bcbe1084bc1d416f41d82ffc0451fbbe1398fa6e479549749051f998a87c",81,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.29.0\u002Fvllm-0.29.0%2Bxpu-cp38-abi3-manylinux_2_34_x86_64.whl",{"fileName":286,"sizeMb":287,"sha256":288,"downloadCount":289,"originUrl":290},"vllm-0.29.0-cp38-abi3-manylinux_2_28_aarch64.whl",295.67,"e6b0dfc2b6fd307315e9b34b73cd2bfe7b6b08958eda721828e61732bba426b0",11330,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.29.0\u002Fvllm-0.29.0-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":292,"sizeMb":293,"sha256":294,"downloadCount":295,"originUrl":296},"vllm-0.29.0-cp38-abi3-manylinux_2_28_x86_64.whl",301.32,"09d48617fc2be9c6cdcd5db480651ab0d84817b257204f2cc2e3ecbb70bbb635",20289,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.29.0\u002Fvllm-0.29.0-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":298,"sizeMb":299,"sha256":300,"downloadCount":301,"originUrl":302},"vllm-0.29.0.tar.gz",39.01,"c03feb07943ecae9e69e7dc424d7c5569ecf8d1292513aef2a4de67af381cd97",252,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.29.0\u002Fvllm-0.29.0.tar.gz",{"tag":304,"publishAt":305,"semVer":306,"prerelease":67,"assets":308},"v0.28.0","2026-08-26T09:46:30Z",[29,307,29],28,[309,315,321,327,333,339,345,351,357],{"fileName":310,"sizeMb":311,"sha256":312,"downloadCount":313,"originUrl":314},"vllm-0.28.0+cpu-cp312-cp312-macosx_11_0_arm64.whl",25.21,"e8c5a3930367b740914a14420efcc3535da2c2dba5bb23d77221ff81094cc630",3924,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.28.0\u002Fvllm-0.28.0%2Bcpu-cp312-cp312-macosx_11_0_arm64.whl",{"fileName":316,"sizeMb":317,"sha256":318,"downloadCount":319,"originUrl":320},"vllm-0.28.0+cpu-cp38-abi3-manylinux_2_34_aarch64.whl",63.1,"38f37515e58a39eb0f23847387ee3790cb5100d3cdfae40a77d0265f135461d1",1592,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.28.0\u002Fvllm-0.28.0%2Bcpu-cp38-abi3-manylinux_2_34_aarch64.whl",{"fileName":322,"sizeMb":323,"sha256":324,"downloadCount":325,"originUrl":326},"vllm-0.28.0+cpu-cp38-abi3-manylinux_2_34_x86_64.whl",124.1,"71f69f1549ba5832595db2e85a485d067ba13b62c60055340128fbda0842bb97",7103,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.28.0\u002Fvllm-0.28.0%2Bcpu-cp38-abi3-manylinux_2_34_x86_64.whl",{"fileName":328,"sizeMb":329,"sha256":330,"downloadCount":331,"originUrl":332},"vllm-0.28.0+cu129-cp38-abi3-manylinux_2_28_aarch64.whl",497,"09ae6e121339b3208b628aa17158bf2300dfbee9658fa296e4565b82afec0511",658,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.28.0\u002Fvllm-0.28.0%2Bcu129-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":334,"sizeMb":335,"sha256":336,"downloadCount":337,"originUrl":338},"vllm-0.28.0+cu129-cp38-abi3-manylinux_2_28_x86_64.whl",522.62,"8ec943b66a0c6b4351d0778e99d7bacfca5788dd8eedd49425092bacb61c4397",26498,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.28.0\u002Fvllm-0.28.0%2Bcu129-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":340,"sizeMb":341,"sha256":342,"downloadCount":343,"originUrl":344},"vllm-0.28.0+xpu-cp38-abi3-manylinux_2_34_x86_64.whl",28.6,"14529b4222c3e1df3a900e41a88ece297cbd656f0a1ea16a993b108496c30eaa",89,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.28.0\u002Fvllm-0.28.0%2Bxpu-cp38-abi3-manylinux_2_34_x86_64.whl",{"fileName":346,"sizeMb":347,"sha256":348,"downloadCount":349,"originUrl":350},"vllm-0.28.0-cp38-abi3-manylinux_2_28_aarch64.whl",294.56,"817b8181f7f61b4a62dc1d5d9ab39f2bfb60a6cb86c29879a78a147b85756787",22114,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.28.0\u002Fvllm-0.28.0-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":352,"sizeMb":353,"sha256":354,"downloadCount":355,"originUrl":356},"vllm-0.28.0-cp38-abi3-manylinux_2_28_x86_64.whl",299.93,"addb0ffdaafd8155d75e9b3f5ddb3da28fdee9e8a7097ede91f7db2e9e1a3889",23090,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.28.0\u002Fvllm-0.28.0-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":358,"sizeMb":359,"sha256":360,"downloadCount":361,"originUrl":362},"vllm-0.28.0.tar.gz",38.14,"ac96dd0ec5be9c13f2aa4bfb50498c4727110406f6e4980b58a4097e4d18634a",434,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.28.0\u002Fvllm-0.28.0.tar.gz",{"tag":364,"publishAt":365,"semVer":366,"prerelease":67,"assets":368},"v0.27.1","2026-08-11T10:47:49Z",[29,367,26],27,[369,375,381,387,393,399,405,411],{"fileName":370,"sizeMb":371,"sha256":372,"downloadCount":373,"originUrl":374},"vllm-0.27.1+cpu-cp312-cp312-macosx_11_0_arm64.whl",24.61,"a3b537892ebc98e97132101aa2096210842b982e05f976fa337caa3f771f9f0c",3044,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.1\u002Fvllm-0.27.1%2Bcpu-cp312-cp312-macosx_11_0_arm64.whl",{"fileName":376,"sizeMb":377,"sha256":378,"downloadCount":379,"originUrl":380},"vllm-0.27.1+cpu-cp38-abi3-manylinux_2_34_aarch64.whl",62.43,"510021af7d5d953040c3b97a112e50902bad7748b238c5d32675c04d23a29e52",5975,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.1\u002Fvllm-0.27.1%2Bcpu-cp38-abi3-manylinux_2_34_aarch64.whl",{"fileName":382,"sizeMb":383,"sha256":384,"downloadCount":385,"originUrl":386},"vllm-0.27.1+cpu-cp38-abi3-manylinux_2_34_x86_64.whl",121.82,"36f0e7b2031233ff09e521716723b0e05ab62054c9a9a05d873af43052140f33",9541,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.1\u002Fvllm-0.27.1%2Bcpu-cp38-abi3-manylinux_2_34_x86_64.whl",{"fileName":388,"sizeMb":389,"sha256":390,"downloadCount":391,"originUrl":392},"vllm-0.27.1+cu129-cp38-abi3-manylinux_2_28_aarch64.whl",494.75,"d73c897b0bde9e515c30101f694989fb65821a8543e7cad8f95e5440dc6bae81",925,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.1\u002Fvllm-0.27.1%2Bcu129-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":394,"sizeMb":395,"sha256":396,"downloadCount":397,"originUrl":398},"vllm-0.27.1+cu129-cp38-abi3-manylinux_2_28_x86_64.whl",520.47,"bf0d52faa2a51e7a01c6856a7a8a2d1307fd0ff711415d34168a67ffac0fa47b",24661,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.1\u002Fvllm-0.27.1%2Bcu129-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":400,"sizeMb":401,"sha256":402,"downloadCount":403,"originUrl":404},"vllm-0.27.1-cp38-abi3-manylinux_2_28_aarch64.whl",292.95,"74c1a04548c9016ace8f6c03592edf62517152a0e833f46282556ffbe0796254",586,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.1\u002Fvllm-0.27.1-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":406,"sizeMb":407,"sha256":408,"downloadCount":409,"originUrl":410},"vllm-0.27.1-cp38-abi3-manylinux_2_28_x86_64.whl",298.39,"98e9fc2a1ed8549a733c9d1b242e2002b82367da9e29e37801761438cb3a2670",5536,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.1\u002Fvllm-0.27.1-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":412,"sizeMb":413,"sha256":414,"downloadCount":415,"originUrl":416},"vllm-0.27.1.tar.gz",37.44,"eec2d54d137ac1e59cb4c39226dfee1943eefc8f4788f5821d7300d6acbdb646",562,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.1\u002Fvllm-0.27.1.tar.gz",{"tag":418,"publishAt":419,"semVer":420,"prerelease":67,"assets":421},"v0.27.0","2026-08-10T21:18:11Z",[29,367,29],[422,427,432,437,442,447,451,456],{"fileName":423,"sizeMb":371,"sha256":424,"downloadCount":425,"originUrl":426},"vllm-0.27.0+cpu-cp312-cp312-macosx_11_0_arm64.whl","72fc5842b13635ce4f80620fc96adfa60d51bbd7f822cf5b5a4d9e9821fd6690",364,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.0\u002Fvllm-0.27.0%2Bcpu-cp312-cp312-macosx_11_0_arm64.whl",{"fileName":428,"sizeMb":377,"sha256":429,"downloadCount":430,"originUrl":431},"vllm-0.27.0+cpu-cp38-abi3-manylinux_2_34_aarch64.whl","9a3e823d214a6ba69fecdf4753623f596007f2f181e4fd275b25426b3b4436b4",26,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.0\u002Fvllm-0.27.0%2Bcpu-cp38-abi3-manylinux_2_34_aarch64.whl",{"fileName":433,"sizeMb":383,"sha256":434,"downloadCount":435,"originUrl":436},"vllm-0.27.0+cpu-cp38-abi3-manylinux_2_34_x86_64.whl","306e4facfb5515de00258d3deacb133eeb2f53f386f7ab2b8b6f49a104e9c703",1286,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.0\u002Fvllm-0.27.0%2Bcpu-cp38-abi3-manylinux_2_34_x86_64.whl",{"fileName":438,"sizeMb":389,"sha256":439,"downloadCount":440,"originUrl":441},"vllm-0.27.0+cu129-cp38-abi3-manylinux_2_28_aarch64.whl","102b4611cd6010fd7ba060db96a83f4d7bf698b8d60b99cce5b9528c421c0edf",95,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.0\u002Fvllm-0.27.0%2Bcu129-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":443,"sizeMb":395,"sha256":444,"downloadCount":445,"originUrl":446},"vllm-0.27.0+cu129-cp38-abi3-manylinux_2_28_x86_64.whl","c1cfee107e76944797b1982502cdf49e22d98b739285ab3971e9d589f841fc1c",1063,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.0\u002Fvllm-0.27.0%2Bcu129-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":448,"sizeMb":401,"sha256":449,"downloadCount":103,"originUrl":450},"vllm-0.27.0-cp38-abi3-manylinux_2_28_aarch64.whl","d38f7f738a6c340fa86abaa56287ab126d0c841f9d129fb6eb21056c3a83e40b","https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.0\u002Fvllm-0.27.0-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":452,"sizeMb":407,"sha256":453,"downloadCount":454,"originUrl":455},"vllm-0.27.0-cp38-abi3-manylinux_2_28_x86_64.whl","02d8265e71bab1cf50f93026211c9a75562a7f0a72b3a32ec5e0e8bcdd62ec75",210,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.0\u002Fvllm-0.27.0-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":457,"sizeMb":458,"sha256":459,"downloadCount":460,"originUrl":461},"vllm-0.27.0.tar.gz",37.43,"862671641625d6cfb7aaab98a5386126738b59ca50e5d01ca539a4e32f63ffcb",65,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.27.0\u002Fvllm-0.27.0.tar.gz",{"tag":463,"publishAt":464,"semVer":465,"prerelease":67,"assets":466},"v0.26.0","2026-07-27T01:06:58Z",[29,430,29],[467,473,479,485,491,497,503,509],{"fileName":468,"sizeMb":469,"sha256":470,"downloadCount":471,"originUrl":472},"vllm-0.26.0+cpu-cp312-cp312-macosx_11_0_arm64.whl",19.77,"c2419fbb091f17915a1e41a952df35d04ecf5373d31c4499bf4a84777e1e8e5a",2989,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.26.0\u002Fvllm-0.26.0%2Bcpu-cp312-cp312-macosx_11_0_arm64.whl",{"fileName":474,"sizeMb":475,"sha256":476,"downloadCount":477,"originUrl":478},"vllm-0.26.0+cpu-cp38-abi3-manylinux_2_34_aarch64.whl",54.79,"aa9500f2026e6b343e4361bdbbe3de10bb357eef32b0815ae9f0c31947d8427d",7589,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.26.0\u002Fvllm-0.26.0%2Bcpu-cp38-abi3-manylinux_2_34_aarch64.whl",{"fileName":480,"sizeMb":481,"sha256":482,"downloadCount":483,"originUrl":484},"vllm-0.26.0+cpu-cp38-abi3-manylinux_2_34_x86_64.whl",111.7,"07de050c09b992c3215615ec11c4bb09eb384760d4e5bbb835496a846b64873d",10331,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.26.0\u002Fvllm-0.26.0%2Bcpu-cp38-abi3-manylinux_2_34_x86_64.whl",{"fileName":486,"sizeMb":487,"sha256":488,"downloadCount":489,"originUrl":490},"vllm-0.26.0+cu129-cp38-abi3-manylinux_2_28_aarch64.whl",472.11,"a21f3a78adffce8f3e88b08ad1145ff5d7d6e0618c652db3e0b914c700cd9b64",70578,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.26.0\u002Fvllm-0.26.0%2Bcu129-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":492,"sizeMb":493,"sha256":494,"downloadCount":495,"originUrl":496},"vllm-0.26.0+cu129-cp38-abi3-manylinux_2_28_x86_64.whl",494.9,"6ce4ca30616f0a35810391015622b197a7b8b267ed27f8716f0789db79ff578b",126114,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.26.0\u002Fvllm-0.26.0%2Bcu129-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":498,"sizeMb":499,"sha256":500,"downloadCount":501,"originUrl":502},"vllm-0.26.0-cp38-abi3-manylinux_2_28_aarch64.whl",284.45,"52a4c3e55c2c80cc8793e52ccc244457ceade25b0ad7caa1c15e5002a95a1b2c",26379,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.26.0\u002Fvllm-0.26.0-cp38-abi3-manylinux_2_28_aarch64.whl",{"fileName":504,"sizeMb":505,"sha256":506,"downloadCount":507,"originUrl":508},"vllm-0.26.0-cp38-abi3-manylinux_2_28_x86_64.whl",289.63,"adb1e4c9b46d0dfdb094121ae5aad670a42412dd813ed4e5db069ed6a15006de",27611,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.26.0\u002Fvllm-0.26.0-cp38-abi3-manylinux_2_28_x86_64.whl",{"fileName":510,"sizeMb":511,"sha256":512,"downloadCount":513,"originUrl":514},"vllm-0.26.0.tar.gz",36.58,"23e9fa19d7e20ce7dcc1c074d41503e2116d23f19e688f5d5ea91b741f958502",676,"https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Freleases\u002Fdownload\u002Fv0.26.0\u002Fvllm-0.26.0.tar.gz",[516,535,564,576,595,610],{"owner":517,"repo":518,"repoId":519,"category":520,"subgroup":61,"language":521,"tags":522,"starCount":524,"ownerUid":525,"latestTag":526,"platforms":527,"lastReleaseTime":532,"downloadCount":533,"description":534},"ggml-org","llama.cpp",612354784,"ai","C++",[523],"ggml",132621,134263123,"b11962",[528,529,530,531],"windows","mac","linux","other","2026-10-28T13:42:20Z",603,"一个用C\u002FC++实现的大语言模型（LLM）推理引擎，支持多种模型格式，提供CLI、服务器和内置Web UI，可在本地运行LLM和VLM。",{"owner":536,"repo":537,"repoId":538,"category":520,"subgroup":61,"language":539,"tags":540,"starCount":557,"ownerUid":558,"latestTag":559,"platforms":560,"lastReleaseTime":561,"downloadCount":562,"description":563},"diegosouzapw","OmniRoute",1157102282,"TypeScript",[541,542,543,544,545,546,547,548,549,550,42,551,552,47,553,554,53,555,55,556],"a2a","ai-agents","ai-gateway","anthropic","claude","claude-code","cline","codex","copilot","cursor","free-ai","gemini","llm-gateway","mcp","openai-proxy","token-saver",78744,8016841,"v3.8.52",[528,529,530,531],"2026-10-19T20:02:45Z",836,"一款免费的 MIT AI 网关，提供统一端点访问 290+ 提供商、500+ 模型，支持智能回退和令牌压缩。",{"owner":565,"repo":566,"repoId":567,"category":520,"subgroup":61,"language":539,"tags":568,"starCount":569,"ownerUid":570,"latestTag":571,"platforms":572,"lastReleaseTime":573,"downloadCount":574,"description":575},"tashfeenahmed","freellmapi",1217040185,[],42483,9307356,"v0.16.11",[528,529,530,531],"2026-10-27T12:29:53Z",368,"一个聚合了34个免费LLM提供商、635个免费模型端点的统一API网关，每月提供约74亿免费令牌，支持智能路由、自动故障转移和加密密钥存储。",{"owner":577,"repo":578,"repoId":579,"category":580,"subgroup":61,"language":581,"tags":582,"starCount":588,"ownerUid":589,"latestTag":590,"platforms":591,"lastReleaseTime":592,"downloadCount":593,"description":594},"AlexsJones","llmfit",1158588224,"tools","Rust",[583,49,584,585,586,587],"gguf","localai","mlx","skill","unsloth",38441,1235925,"v1.1.18",[528,529,530],"2026-10-28T08:33:20Z",145,"一个帮助用户根据硬件配置推荐最佳本地大语言模型的命令行工具，支持多GPU、MoE架构和多种量化格式。",{"owner":596,"repo":597,"repoId":598,"category":520,"subgroup":61,"language":599,"tags":600,"starCount":603,"ownerUid":604,"latestTag":605,"platforms":606,"lastReleaseTime":607,"downloadCount":608,"description":609},"yetone","magpie",1383809012,"Go",[546,548,42,601,49,602],"gemini-cli","macos",24427,1206493,"v0.1.1906",[528,529],"2026-10-28T13:48:58Z",14,"Magpie 是一个本地运行的 AI 网关，让 Claude Code、Codex、Gemini CLI、OpenCode 等编程代理统一使用 Kimi、DeepSeek、GLM、ChatGPT 等模型，菜单栏一键切换，配额用尽自动切换下一账号。",{"owner":611,"repo":612,"repoId":613,"category":520,"subgroup":61,"language":19,"tags":614,"starCount":618,"ownerUid":619,"latestTag":620,"platforms":621,"lastReleaseTime":622,"downloadCount":403,"description":623},"jundot","omlx",1157171418,[615,616,49,602,585,617],"apple-silicon","inference-server","openai-api",23094,64250138,"v0.7.3rc1",[529,531],"2026-10-27T09:22:10Z","一个专为Apple Silicon优化的LLM推理服务器，支持连续批处理和SSD缓存，通过菜单栏进行管理。",{"description":625,"readmeHtml":626},"vLLM 是一个高吞吐、省显存的大语言模型推理与服务引擎（Python），以 PagedAttention 与连续批处理为核心，提供 OpenAI 兼容 API，支持 200 多种模型架构与多种硬件后端。","\u003Ch2 id=\"项目概述\">项目概述\u003C\u002Fh2>\u003Cp>vLLM 是一个快速易用的大语言模型推理与服务库，诞生于 UC Berkeley，现由两千多名贡献者共同维护。它通过 PagedAttention 管理 GPU 显存、以连续批处理调度请求，能在相同硬件上大幅提升服务吞吐，并直接提供 OpenAI 兼容 API。支持 200 多种 Hugging Face 模型架构，覆盖 LLM、多模态与嵌入模型。\u003C\u002Fp>\n\u003Cblockquote>\n\u003Cp>一条命令把大模型变成高性能 API 服务，硬件利用率显著提升。\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\u003Ctable>\n\u003Cthead>\n\u003Ctr>\n\u003Cth>项目\u003C\u002Fth>\n\u003Cth>内容\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\u003Ctr>\n\u003Ctd>平台\u003C\u002Ftd>\n\u003Ctd>Linux 为主，支持 NVIDIA、AMD、Intel GPU 与 x86\u002FARM CPU\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>语言\u003C\u002Ftd>\n\u003Ctd>Python\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>许可\u003C\u002Ftd>\n\u003Ctd>Apache-2.0\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>模型支持\u003C\u002Ftd>\n\u003Ctd>200+ Hugging Face 架构\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\u003C\u002Ftable>\n\u003Ch2 id=\"主要功能\">主要功能\u003C\u002Fh2>\u003Cul>\n\u003Cli>\u003Cstrong>PagedAttention\u003C\u002Fstrong>：像管理操作系统内存页一样管理 KV Cache，显存利用率高\u003C\u002Fli>\n\u003Cli>\u003Cstrong>连续批处理\u003C\u002Fstrong>：请求随到随入批，配合分块预填充与前缀缓存\u003C\u002Fli>\n\u003Cli>\u003Cstrong>高性能内核\u003C\u002Fstrong>：内置 FlashAttention、FlashInfer 等优化注意力内核\u003C\u002Fli>\n\u003Cli>\u003Cstrong>量化支持\u003C\u002Fstrong>：FP8、INT8、GPTQ、AWQ、GGUF 等多种量化方案\u003C\u002Fli>\n\u003Cli>\u003Cstrong>OpenAI 兼容 API\u003C\u002Fstrong>：直接对接现有 OpenAI SDK 生态，另有 Anthropic 与 gRPC 接口\u003C\u002Fli>\n\u003Cli>\u003Cstrong>结构化输出\u003C\u002Fstrong>：用 xgrammar 等约束生成 JSON 等固定格式\u003C\u002Fli>\n\u003Cli>\u003Cstrong>分布式推理\u003C\u002Fstrong>：张量、流水线、数据、专家并行多种切分方式\u003C\u002Fli>\n\u003Cli>\u003Cstrong>多 LoRA\u003C\u002Fstrong>：一个服务实例同时挂载多个 LoRA 适配器\u003C\u002Fli>\n\u003Cli>\u003Cstrong>投机解码\u003C\u002Fstrong>：n-gram、EAGLE 等方案加速生成\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2 id=\"适用人群\">适用人群\u003C\u002Fh2>\u003Cul>\n\u003Cli>自托管大模型 API 的后端工程师\u003C\u002Fli>\n\u003Cli>需要批量离线推理的数据团队\u003C\u002Fli>\n\u003Cli>把开源模型部署到生产环境的企业\u003C\u002Fli>\n\u003Cli>研究 LLM 服务性能的算法工程师\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2 id=\"快速上手\">快速上手\u003C\u002Fh2>\u003Col>\n\u003Cli>安装（推荐 uv）：\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Cdiv class=\"code-block\">\u003Cpre>\u003Ccode class=\"hljs language-bash\">uv pip install vllm\nvllm download-kernels\u003C\u002Fcode>\u003C\u002Fpre>\u003C\u002Fdiv>\u003Col start=\"2\">\n\u003Cli>启动 OpenAI 兼容服务：\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Cdiv class=\"code-block\">\u003Cpre>\u003Ccode class=\"hljs language-bash\">vllm serve &lt;模型名&gt;\u003C\u002Fcode>\u003C\u002Fpre>\u003C\u002Fdiv>\u003Col start=\"3\">\n\u003Cli>把 OpenAI SDK 指向本地端口即可调用，详见官方快速开始文档\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Ch2 id=\"核心优势\">核心优势\u003C\u002Fh2>\u003Cul>\n\u003Cli>\u003Cstrong>吞吐高\u003C\u002Fstrong>：同等硬件下服务能力大幅提升，成本随之降低\u003C\u002Fli>\n\u003Cli>\u003Cstrong>生态兼容\u003C\u002Fstrong>：Hugging Face 模型即插即用，API 与 OpenAI 一致\u003C\u002Fli>\n\u003Cli>\u003Cstrong>硬件面广\u003C\u002Fstrong>：NVIDIA、AMD、Intel GPU 与 CPU、TPU 等多种后端\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2 id=\"系统要求\">系统要求\u003C\u002Fh2>\u003Ctable>\n\u003Cthead>\n\u003Ctr>\n\u003Cth>项目\u003C\u002Fth>\n\u003Cth>要求\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\u003Ctr>\n\u003Ctd>操作系统\u003C\u002Ftd>\n\u003Ctd>Linux 为主\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>硬件\u003C\u002Ftd>\n\u003Ctd>NVIDIA\u002FAMD\u002FIntel GPU、x86\u002FARM CPU、TPU 等\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>语言环境\u003C\u002Ftd>\n\u003Ctd>Python\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\u003C\u002Ftable>\n\u003Ch2 id=\"注意事项\">注意事项\u003C\u002Fh2>\u003Cul>\n\u003Cli>\u003Cstrong>安装环境\u003C\u002Fstrong>：GPU 版需按本机 CUDA 环境选择对应 PyTorch wheel，从源码编译更复杂\u003C\u002Fli>\n\u003Cli>\u003Cstrong>内核预下载\u003C\u002Fstrong>：安装后执行 \u003Ccode>vllm download-kernels\u003C\u002Fcode> 可加快服务启动\u003C\u002Fli>\n\u003C\u002Ful>\n",1793197965978]