[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skills:sce6":3,"skills:pro6":50,"skills:creators9":95,"skills:stats":193,"skills:all:stars:26764:":195},[4,15,22,29,36,43],{"categoryNo":5,"taxonomyType":6,"taxonomyTypeDesc":7,"categoryName":8,"categoryNameEn":9,"description":10,"level":11,"skillCount":12,"parentNo":13,"parentName":13,"children":14},"01JBSEEDCAT000000000000022","SCENARIO","使用场景","智能体与自动化","Agents & Automation","Skill、MCP、智能体与工作流自动化",1,39675,null,[],{"categoryNo":16,"taxonomyType":6,"taxonomyTypeDesc":7,"categoryName":17,"categoryNameEn":18,"description":19,"level":11,"skillCount":20,"parentNo":13,"parentName":13,"children":21},"01JBSEEDCAT000000000000006","软件开发","Software Development","编码、测试、部署等软件研发任务",26798,[],{"categoryNo":23,"taxonomyType":6,"taxonomyTypeDesc":7,"categoryName":24,"categoryNameEn":25,"description":26,"level":11,"skillCount":27,"parentNo":13,"parentName":13,"children":28},"01JBSEEDCAT000000000000014","设计创意","Design & Creative","视觉设计、品牌与多媒体创作",13847,[],{"categoryNo":30,"taxonomyType":6,"taxonomyTypeDesc":7,"categoryName":31,"categoryNameEn":32,"description":33,"level":11,"skillCount":34,"parentNo":13,"parentName":13,"children":35},"01JBSEEDCAT000000000000001","办公效率","Office Productivity","日常办公文档、表格、演示与沟通类任务",6403,[],{"categoryNo":37,"taxonomyType":6,"taxonomyTypeDesc":7,"categoryName":38,"categoryNameEn":39,"description":40,"level":11,"skillCount":41,"parentNo":13,"parentName":13,"children":42},"01JBSEEDCAT000000000000017","内容创作","Content Creation","写作、文案与营销内容生产",6063,[],{"categoryNo":44,"taxonomyType":6,"taxonomyTypeDesc":7,"categoryName":45,"categoryNameEn":46,"description":47,"level":11,"skillCount":48,"parentNo":13,"parentName":13,"children":49},"01JBSEEDCAT000000000000020","研究与学习","Research & Learning","资料检索、知识整理与学习辅导",5574,[],[51,60,67,74,81,88],{"categoryNo":52,"taxonomyType":53,"taxonomyTypeDesc":54,"categoryName":55,"categoryNameEn":56,"description":57,"level":11,"skillCount":58,"parentNo":13,"parentName":13,"children":59},"01JBSEEDCAT000000000000030","PROFESSION","适用职业","开发工程师","Software Engineer","从事软件编码、测试与运维的工程师",13341,[],{"categoryNo":61,"taxonomyType":53,"taxonomyTypeDesc":54,"categoryName":62,"categoryNameEn":63,"description":64,"level":11,"skillCount":65,"parentNo":13,"parentName":13,"children":66},"01JBSEEDCAT000000000000032","设计师","Designer","视觉、交互与品牌设计人员",7902,[],{"categoryNo":68,"taxonomyType":53,"taxonomyTypeDesc":54,"categoryName":69,"categoryNameEn":70,"description":71,"level":11,"skillCount":72,"parentNo":13,"parentName":13,"children":73},"01JBSEEDCAT000000000000033","数据分析师","Data Analyst","数据分析、报表与商业智能人员",5005,[],{"categoryNo":75,"taxonomyType":53,"taxonomyTypeDesc":54,"categoryName":76,"categoryNameEn":77,"description":78,"level":11,"skillCount":79,"parentNo":13,"parentName":13,"children":80},"01JBSEEDCAT000000000000036","内容创作者","Content Creator","作者、编辑与自媒体创作者",3978,[],{"categoryNo":82,"taxonomyType":53,"taxonomyTypeDesc":54,"categoryName":83,"categoryNameEn":84,"description":85,"level":11,"skillCount":86,"parentNo":13,"parentName":13,"children":87},"01JBSEEDCAT000000000000035","市场营销","Marketer","市场推广、增长与品牌运营人员",3454,[],{"categoryNo":89,"taxonomyType":53,"taxonomyTypeDesc":54,"categoryName":90,"categoryNameEn":91,"description":92,"level":11,"skillCount":93,"parentNo":13,"parentName":13,"children":94},"01JBSEEDCAT000000000000037","律师与法务","Legal","合同、合规与法律事务人员",3337,[],{"records":96,"total":191,"current":11,"size":192},[97,111,120,129,139,149,159,169,179],{"accountNo":98,"login":99,"displayName":100,"avatarUrl":101,"accountType":102,"accountTypeDesc":103,"profileDescription":104,"skillCount":105,"repoCount":106,"professionCount":107,"scenarioCount":108,"totalStars":109,"updatedAt":110},"01M698VYDXZ6JW5WH3CRV650R8","microsoft","Microsoft","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F6154722?v=4","ORGANIZATION","组织","Open source projects and samples from Microsoft",4300,236,11,7,832929,"2026-11-15 04:18:35",{"accountNo":112,"login":113,"displayName":113,"avatarUrl":114,"accountType":102,"accountTypeDesc":103,"profileDescription":115,"skillCount":116,"repoCount":117,"professionCount":108,"scenarioCount":108,"totalStars":118,"updatedAt":119},"01M698P3FSAZ6B4R928ZF19KPT","openclaw","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F252820863?v=4","Your personal, open source AI assistant.",369,32,437811,"2026-11-15 03:04:02",{"accountNo":121,"login":122,"displayName":123,"avatarUrl":124,"accountType":102,"accountTypeDesc":103,"profileDescription":13,"skillCount":125,"repoCount":126,"professionCount":107,"scenarioCount":108,"totalStars":127,"updatedAt":128},"01M697XGDPXJAVQ8391V2BCYTE","anthropics","Anthropic","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F76263028?v=4",716,18,427323,"2026-11-15 05:52:49",{"accountNo":130,"login":131,"displayName":132,"avatarUrl":133,"accountType":102,"accountTypeDesc":103,"profileDescription":13,"skillCount":134,"repoCount":135,"professionCount":108,"scenarioCount":136,"totalStars":137,"updatedAt":138},"01M6AC29YGCR4S2NTJFV0G8B81","openai","OpenAI","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F14957082?v=4",177,23,6,344400,"2026-11-15 03:07:28",{"accountNo":140,"login":141,"displayName":142,"avatarUrl":143,"accountType":102,"accountTypeDesc":103,"profileDescription":13,"skillCount":144,"repoCount":145,"professionCount":146,"scenarioCount":108,"totalStars":147,"updatedAt":148},"01M6AC83E6K4FDKK7HZ8WNMQKE","apache","The Apache Software Foundation","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F47359?v=4",440,76,8,298252,"2026-11-15 04:11:16",{"accountNo":150,"login":151,"displayName":152,"avatarUrl":153,"accountType":102,"accountTypeDesc":103,"profileDescription":154,"skillCount":155,"repoCount":156,"professionCount":107,"scenarioCount":108,"totalStars":157,"updatedAt":158},"01M6AC83DS133TPXFSP1DHNCC6","github","GitHub","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F9919?v=4","How people build software.",1022,30,278314,"2026-11-15 06:45:15",{"accountNo":160,"login":161,"displayName":162,"avatarUrl":163,"accountType":102,"accountTypeDesc":103,"profileDescription":164,"skillCount":165,"repoCount":166,"professionCount":107,"scenarioCount":108,"totalStars":167,"updatedAt":168},"01M698VYEFDM0VJBZDYYJ2T83S","google","Google","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F1342004?v=4","Google ❤️ Open Source",542,50,243804,"2026-11-15 03:00:30",{"accountNo":170,"login":171,"displayName":172,"avatarUrl":173,"accountType":102,"accountTypeDesc":103,"profileDescription":13,"skillCount":174,"repoCount":175,"professionCount":176,"scenarioCount":108,"totalStars":177,"updatedAt":178},"01M6AC83EPDGBFN19MWYRREJD0","Significant-Gravitas","AutoGPT","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F130738209?v=4",87,2,10,188029,"2026-11-15 06:21:09",{"accountNo":180,"login":181,"displayName":182,"avatarUrl":183,"accountType":184,"accountTypeDesc":185,"profileDescription":186,"skillCount":187,"repoCount":188,"professionCount":136,"scenarioCount":108,"totalStars":189,"updatedAt":190},"01M6DJCCDDG53MTYYRTXAXFG17","garrytan","Garry Tan","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F19957?v=4","USER","个人作者","Writes software, dreams. ",340,3,175019,"2026-11-15 03:57:03",5953,9,{"skillCount":194,"professionCount":107,"scenarioCount":108,"creatorCount":191},482556,{"records":196,"total":194,"current":454,"size":126},[197,237,264,270,289,294,319,324,345,350,371,376,397,402,416,421,437,442],{"skillNo":198,"name":199,"nameZh":13,"description":200,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":208,"categories":209,"keywords":213,"publishedAt":236},"01M6KQ3KN9XBCK6VPRGEMY8HVB","langsmith-observability","LLM observability platform for tracing, evaluation, and monitoring. Use when debugging LLM applications, evaluating model outputs against datasets, monitoring production systems, or building systematic testing pipelines for AI applications.","endman100\u002Fskill-router-nexus","endman100","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F31997019?v=4","01M6KPDX17BYDWW4RRB0TNGGMX","01M6KPDX64Y2QBEWVFG5W5PWPF",0,"MIT","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-MLOps\u002Flangsmith\u002FSKILL.md",[210],{"categoryNo":211,"categoryName":212,"taxonomyType":6,"taxonomyTypeDesc":7},"01JBSEEDCAT000000000000008","测试与质量",[214,215,216,217,218,219,220,221,222,223,224,225,226,227,228,229,230,231,232,233,234,235],"observability","LangSmith","langsmith observability","testing","research","debugging","tracing","monitoring","Evaluation","applications","model","ops","use","datasets","observability platform","platform","systematic","LLM Applications","orchestra","for","outputs","when","2026-11-10 10:12:30",{"skillNo":238,"name":239,"nameZh":13,"description":240,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":241,"categories":242,"keywords":246,"publishedAt":236},"01M6KQ3DPREJMMVHZBS7MXFDJ1","serving-llms-vllm","Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency\u002Fthroughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ\u002FAWQ\u002FFP8), and tensor parallelism.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fvllm\u002Fvllm\u002FSKILL.md",[243],{"categoryNo":244,"categoryName":245,"taxonomyType":6,"taxonomyTypeDesc":7},"01JBSEEDCAT000000000000010","后端与接口",[247,248,249,250,251,131,252,253,254,255,256,226,257,258,259,260,261,262,232,263,235],"llms","vLLM","serving llms vllm","serving","memory","openai api","FP8","Inference Serving","Transformers","api","latency","AWQ","GPTQ","throughput","with","openai-compatible","batching",{"skillNo":265,"name":239,"nameZh":13,"description":240,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":266,"categories":267,"keywords":269,"publishedAt":236},"01M6KQ38RR2WHG23XVQFDW7CQS","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fvllm\u002FSKILL.md",[268],{"categoryNo":244,"categoryName":245,"taxonomyType":6,"taxonomyTypeDesc":7},[247,248,249,250,251,131,252,253,254,255,256,226,257,258,259,260,261,262,232,263,235],{"skillNo":271,"name":272,"nameZh":13,"description":273,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":274,"categories":275,"keywords":276,"publishedAt":236},"01M6KQ34SK78ZHBZM1AWNPJQ2Z","tensorrt-llm","Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100\u002FH100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8\u002FINT4), in-flight batching, and multi-GPU scaling.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Ftensorrt-llm\u002Ftensorrt-llm\u002FSKILL.md",[],[277,278,279,280,281,253,282,254,283,226,284,257,285,286,287,260,261,288,232,233,263,235],"tensorrt llm","tensorrt","pytorch","deployment","Optimization","NVIDIA","llm inference","scaling","TensorRT-LLM","multi","H100","you",{"skillNo":290,"name":272,"nameZh":13,"description":273,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":291,"categories":292,"keywords":293,"publishedAt":236},"01M6KQ30X8607AHGHWM5EDMV9A","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Ftensorrt-llm\u002FSKILL.md",[],[277,278,279,280,281,253,282,254,283,226,284,257,285,286,287,260,261,288,232,233,263,235],{"skillNo":295,"name":296,"nameZh":13,"description":297,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":298,"categories":299,"keywords":300,"publishedAt":236},"01M6KQ2TTTBX2X7RK9FKNVC40N","speculative-decoding","Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques. Use when optimizing inference speed (1.5-3.6× speedup), reducing latency for real-time applications, or deploying models with limited compute. Covers draft models, tree-based attention, Jacobi iteration, parallel token generation, and production deployment strategies.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fspeculative-decoding\u002Fspeculative-decoding\u002FSKILL.md",[],[301,302,255,303,304,305,306,307,308,309,257,310,311,312,313,314,315,316,317,318,235],"speculative decoding","Fast Inference","parallel","deployment strategies","time","strategies","attention","token","iteration","Accelerate","Emerging Techniques","medusa","Lookahead Decoding","Draft Models","Tree Attention","Parallel Generation","Latency Reduction","tree",{"skillNo":320,"name":296,"nameZh":13,"description":297,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":321,"categories":322,"keywords":323,"publishedAt":236},"01M6KQ2NXC3KJS8MJ51CY9HCY3","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fspeculative-decoding\u002FSKILL.md",[],[301,302,255,303,304,305,306,307,308,309,257,310,311,312,313,314,315,316,317,318,235],{"skillNo":325,"name":326,"nameZh":13,"description":327,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":328,"categories":329,"keywords":330,"publishedAt":236},"01M6KQ2H673YNBBCWR9YTK3XN4","sglang","Fast structured generation and serving for LLMs with RadixAttention prefix caching. Use for JSON\u002Fregex outputs, constrained decoding, agentic workflows with tool calls, or when you need 5× faster inference than vLLM with prefix sharing. Powers 300,000+ GPUs at xAI, AMD, NVIDIA, and LinkedIn.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fsglang\u002Fsglang\u002FSKILL.md",[],[331,332,247,333,334,282,254,335,336,337,338,339,302,340,255,341,342,343,288,344,235],"SGLang","agentic","workflows","agentic workflows","Structured Generation","RadixAttention","Prefix Caching","Constrained Decoding","JSON Output","Production Scale","xai","caching","regex","at",{"skillNo":346,"name":326,"nameZh":13,"description":327,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":347,"categories":348,"keywords":349,"publishedAt":236},"01M6KQ2CQNSSM76QDB2MT9NQCN","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fsglang\u002FSKILL.md",[],[331,332,247,333,334,282,254,335,336,337,338,339,302,340,255,341,342,343,288,344,235],{"skillNo":351,"name":352,"nameZh":13,"description":353,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":354,"categories":355,"keywords":356,"publishedAt":236},"01M6KQ280E8GB40QZF2WCHT3X8","model-pruning","Reduce LLM size and accelerate inference using pruning techniques like Wanda and SparseGPT. Use when compressing models without retraining, achieving 50% sparsity with minimal accuracy loss, or enabling faster inference on hardware accelerators. Covers unstructured pruning, structured pruning, N:M sparsity, magnitude pruning, and one-shot methods.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fmodel-pruning\u002Fmodel-pruning\u002FSKILL.md",[],[224,357,302,255,358,310,359,311,360,361,362,363,364,365,366,367,368,232,369,370,235],"Model Pruning","methods","Model Compression","Wanda","SparseGPT","Sparsity","N:M Sparsity","One-Shot Pruning","Structured Pruning","Unstructured Pruning","unstructured","techniques","reduce","accuracy",{"skillNo":372,"name":352,"nameZh":13,"description":353,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":373,"categories":374,"keywords":375,"publishedAt":236},"01M6KQ23BGG9JP587B4PMHBYMD","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fmodel-pruning\u002FSKILL.md",[],[224,357,302,255,358,310,359,311,360,361,362,363,364,365,366,367,368,232,369,370,235],{"skillNo":377,"name":378,"nameZh":13,"description":379,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":380,"categories":381,"keywords":382,"publishedAt":236},"01M6KQ1Z8DNFE4JRX5BCAYJXV6","llama-cpp","Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1\u002FM2\u002FM3 Macs, AMD\u002FIntel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fllama-cpp\u002Fllama-cpp\u002FSKILL.md",[],[383,384,385,251,279,282,254,386,387,388,389,390,391,392,393,394,395,396,232,233,235],"cpp","Llama","llama cpp","apple","bit","silicon","CPU","GGUF","llama.cpp","CPU Inference","Edge Deployment","Non-NVIDIA","AMD GPUs","Intel GPUs",{"skillNo":398,"name":378,"nameZh":13,"description":379,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":399,"categories":400,"keywords":401,"publishedAt":236},"01M6KQ1V8C4MCT1HDM2JNPER05","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fllama-cpp\u002FSKILL.md",[],[383,384,385,251,279,282,254,386,387,388,389,390,391,392,393,394,395,396,232,233,235],{"skillNo":403,"name":404,"nameZh":13,"description":405,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":406,"categories":407,"keywords":408,"publishedAt":236},"01M6KQ1PWEZJPVP1K43K1565E1","hqq-quantization","Half-Quadratic Quantization for LLMs without calibration data. Use when quantizing models to 4\u002F3\u002F2-bit precision without needing calibration datasets, for fast quantization workflows, or when deploying with vLLM or HuggingFace Transformers.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fhqq\u002Fhqq\u002FSKILL.md",[],[409,410,411,247,251,333,281,412,248,255,224,226,413,387,414,359,261,232,233,415,235],"Quantization","HQQ","hqq quantization","data","efficiency","Memory Efficiency","calibration",{"skillNo":417,"name":404,"nameZh":13,"description":405,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":418,"categories":419,"keywords":420,"publishedAt":236},"01M6KQ1J6AJWBQ2TAVZZWS4HFJ","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fhqq\u002FSKILL.md",[],[409,410,411,247,251,333,281,412,248,255,224,226,413,387,414,359,261,232,233,415,235],{"skillNo":422,"name":423,"nameZh":13,"description":424,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":425,"categories":426,"keywords":427,"publishedAt":236},"01M6KQ1CT6JXE4NDFSQJ0MX30F","gptq","Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with \u003C2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fgptq\u002Fgptq\u002FSKILL.md",[],[259,247,251,428,281,429,302,430,255,431,432,387,433,434,435,436,288,232,233,370,235],"training","FP16","perplexity","tuning","auto","4-Bit","Memory Optimization","Consumer GPUs","Group-Wise Quantization",{"skillNo":438,"name":423,"nameZh":13,"description":424,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":439,"categories":440,"keywords":441,"publishedAt":236},"01M6KQ17X75YY66NBWNWFD6KEE","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fgptq\u002FSKILL.md",[],[259,247,251,428,281,429,302,430,255,431,432,387,433,434,435,436,288,232,233,370,235],{"skillNo":443,"name":444,"nameZh":13,"description":445,"descriptionZh":13,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":13,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":13,"sourceHtmlUrl":446,"categories":447,"keywords":448,"publishedAt":236},"01M6KQ13JJ9HRV8VASVNXP58CD","gguf-quantization","GGUF format and llama.cpp quantization for efficient CPU\u002FGPU inference. Use when deploying models on consumer hardware, Apple Silicon, or when needing flexible quantization from 2-8 bit without GPU requirements.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fgguf\u002Fgguf\u002FSKILL.md",[],[449,409,390,450,383,451,281,452,224,226,453,386,387,388,389,391,392,359,232,233,235],"gguf quantization","python","gpu","apple silicon","requirements",26764]