[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skills:stats":3,"skills:sce6":8,"skills:pro6":55,"skills:creators9":100,"skills:all:stars:10879:":195},{"skillCount":4,"professionCount":5,"scenarioCount":6,"creatorCount":7},196628,11,7,1905,[9,20,27,34,41,48],{"categoryNo":10,"taxonomyType":11,"taxonomyTypeDesc":12,"categoryName":13,"categoryNameEn":14,"description":15,"level":16,"skillCount":17,"parentNo":18,"parentName":18,"children":19},"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",1,39462,null,[],{"categoryNo":21,"taxonomyType":11,"taxonomyTypeDesc":12,"categoryName":22,"categoryNameEn":23,"description":24,"level":16,"skillCount":25,"parentNo":18,"parentName":18,"children":26},"01JBSEEDCAT000000000000006","软件开发","Software Development","编码、测试、部署等软件研发任务",26546,[],{"categoryNo":28,"taxonomyType":11,"taxonomyTypeDesc":12,"categoryName":29,"categoryNameEn":30,"description":31,"level":16,"skillCount":32,"parentNo":18,"parentName":18,"children":33},"01JBSEEDCAT000000000000014","设计创意","Design & Creative","视觉设计、品牌与多媒体创作",13572,[],{"categoryNo":35,"taxonomyType":11,"taxonomyTypeDesc":12,"categoryName":36,"categoryNameEn":37,"description":38,"level":16,"skillCount":39,"parentNo":18,"parentName":18,"children":40},"01JBSEEDCAT000000000000001","办公效率","Office Productivity","日常办公文档、表格、演示与沟通类任务",6291,[],{"categoryNo":42,"taxonomyType":11,"taxonomyTypeDesc":12,"categoryName":43,"categoryNameEn":44,"description":45,"level":16,"skillCount":46,"parentNo":18,"parentName":18,"children":47},"01JBSEEDCAT000000000000017","内容创作","Content Creation","写作、文案与营销内容生产",6033,[],{"categoryNo":49,"taxonomyType":11,"taxonomyTypeDesc":12,"categoryName":50,"categoryNameEn":51,"description":52,"level":16,"skillCount":53,"parentNo":18,"parentName":18,"children":54},"01JBSEEDCAT000000000000020","研究与学习","Research & Learning","资料检索、知识整理与学习辅导",5531,[],[56,65,72,79,86,93],{"categoryNo":57,"taxonomyType":58,"taxonomyTypeDesc":59,"categoryName":60,"categoryNameEn":61,"description":62,"level":16,"skillCount":63,"parentNo":18,"parentName":18,"children":64},"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",13276,[],{"categoryNo":66,"taxonomyType":58,"taxonomyTypeDesc":59,"categoryName":67,"categoryNameEn":68,"description":69,"level":16,"skillCount":70,"parentNo":18,"parentName":18,"children":71},"01JBSEEDCAT000000000000032","设计师","Designer","视觉、交互与品牌设计人员",7776,[],{"categoryNo":73,"taxonomyType":58,"taxonomyTypeDesc":59,"categoryName":74,"categoryNameEn":75,"description":76,"level":16,"skillCount":77,"parentNo":18,"parentName":18,"children":78},"01JBSEEDCAT000000000000033","数据分析师","Data Analyst","数据分析、报表与商业智能人员",4988,[],{"categoryNo":80,"taxonomyType":58,"taxonomyTypeDesc":59,"categoryName":81,"categoryNameEn":82,"description":83,"level":16,"skillCount":84,"parentNo":18,"parentName":18,"children":85},"01JBSEEDCAT000000000000036","内容创作者","Content Creator","作者、编辑与自媒体创作者",3940,[],{"categoryNo":87,"taxonomyType":58,"taxonomyTypeDesc":59,"categoryName":88,"categoryNameEn":89,"description":90,"level":16,"skillCount":91,"parentNo":18,"parentName":18,"children":92},"01JBSEEDCAT000000000000035","市场营销","Marketer","市场推广、增长与品牌运营人员",3449,[],{"categoryNo":94,"taxonomyType":58,"taxonomyTypeDesc":59,"categoryName":95,"categoryNameEn":96,"description":97,"level":16,"skillCount":98,"parentNo":18,"parentName":18,"children":99},"01JBSEEDCAT000000000000037","律师与法务","Legal","合同、合规与法律事务人员",3317,[],{"records":101,"total":7,"current":16,"size":194},[102,114,123,133,143,153,163,172,182],{"accountNo":103,"login":104,"displayName":105,"avatarUrl":106,"accountType":107,"accountTypeDesc":108,"profileDescription":109,"skillCount":110,"repoCount":111,"professionCount":5,"scenarioCount":6,"totalStars":112,"updatedAt":113},"01M698VYDXZ6JW5WH3CRV650R8","microsoft","Microsoft","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F6154722?v=4","ORGANIZATION","组织","Open source projects and samples from Microsoft",4166,229,828075,"2026-11-12 04:48:47",{"accountNo":115,"login":116,"displayName":116,"avatarUrl":117,"accountType":107,"accountTypeDesc":108,"profileDescription":118,"skillCount":119,"repoCount":120,"professionCount":6,"scenarioCount":6,"totalStars":121,"updatedAt":122},"01M698P3FSAZ6B4R928ZF19KPT","openclaw","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F252820863?v=4","Your personal, open source AI assistant.",362,32,437637,"2026-11-12 03:04:11",{"accountNo":124,"login":125,"displayName":126,"avatarUrl":127,"accountType":107,"accountTypeDesc":108,"profileDescription":18,"skillCount":128,"repoCount":129,"professionCount":6,"scenarioCount":130,"totalStars":131,"updatedAt":132},"01M6AC29YGCR4S2NTJFV0G8B81","openai","OpenAI","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F14957082?v=4",171,23,6,341786,"2026-11-12 03:20:15",{"accountNo":134,"login":135,"displayName":136,"avatarUrl":137,"accountType":107,"accountTypeDesc":108,"profileDescription":18,"skillCount":138,"repoCount":139,"professionCount":140,"scenarioCount":6,"totalStars":141,"updatedAt":142},"01M6AC83E6K4FDKK7HZ8WNMQKE","apache","The Apache Software Foundation","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F47359?v=4",420,75,8,298091,"2026-11-12 05:32:19",{"accountNo":144,"login":145,"displayName":146,"avatarUrl":147,"accountType":107,"accountTypeDesc":108,"profileDescription":148,"skillCount":149,"repoCount":150,"professionCount":5,"scenarioCount":6,"totalStars":151,"updatedAt":152},"01M698VYEFDM0VJBZDYYJ2T83S","google","Google","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F1342004?v=4","Google ❤️ Open Source",535,48,241637,"2026-11-12 03:57:04",{"accountNo":154,"login":155,"displayName":156,"avatarUrl":157,"accountType":107,"accountTypeDesc":108,"profileDescription":158,"skillCount":159,"repoCount":160,"professionCount":5,"scenarioCount":6,"totalStars":161,"updatedAt":162},"01M6AC83DS133TPXFSP1DHNCC6","github","GitHub","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F9919?v=4","How people build software.",936,27,238133,"2026-11-12 03:23:28",{"accountNo":164,"login":165,"displayName":166,"avatarUrl":167,"accountType":107,"accountTypeDesc":108,"profileDescription":18,"skillCount":168,"repoCount":169,"professionCount":5,"scenarioCount":6,"totalStars":170,"updatedAt":171},"01M697XGDPXJAVQ8391V2BCYTE","anthropics","Anthropic","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F76263028?v=4",674,17,231455,"2026-11-12 03:40:35",{"accountNo":173,"login":174,"displayName":175,"avatarUrl":176,"accountType":107,"accountTypeDesc":108,"profileDescription":18,"skillCount":177,"repoCount":178,"professionCount":179,"scenarioCount":6,"totalStars":180,"updatedAt":181},"01M6AC83EPDGBFN19MWYRREJD0","Significant-Gravitas","AutoGPT","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F130738209?v=4",86,2,10,187999,"2026-11-12 03:17:01",{"accountNo":183,"login":184,"displayName":185,"avatarUrl":186,"accountType":187,"accountTypeDesc":188,"profileDescription":189,"skillCount":190,"repoCount":191,"professionCount":130,"scenarioCount":6,"totalStars":192,"updatedAt":193},"01M6DJCCDDG53MTYYRTXAXFG17","garrytan","Garry Tan","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F19957?v=4","USER","个人作者","Writes software, dreams. ",324,3,174404,"2026-11-12 05:14:59",9,{"records":196,"total":4,"current":448,"size":449},[197,237,243,270,276,295,300,325,330,351,356,377,382,403,408,422,427,443],{"skillNo":198,"name":199,"nameZh":18,"description":200,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":208,"categories":209,"keywords":213,"publishedAt":236},"01M6KQ3R3DM267BTEPYGWYGC5N","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\u002Flangsmith\u002FSKILL.md",[210],{"categoryNo":211,"categoryName":212,"taxonomyType":11,"taxonomyTypeDesc":12},"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":199,"nameZh":18,"description":200,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":239,"categories":240,"keywords":242,"publishedAt":236},"01M6KQ3KN9XBCK6VPRGEMY8HVB","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-MLOps\u002Flangsmith\u002FSKILL.md",[241],{"categoryNo":211,"categoryName":212,"taxonomyType":11,"taxonomyTypeDesc":12},[214,215,216,217,218,219,220,221,222,223,224,225,226,227,228,229,230,231,232,233,234,235],{"skillNo":244,"name":245,"nameZh":18,"description":246,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":247,"categories":248,"keywords":252,"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",[249],{"categoryNo":250,"categoryName":251,"taxonomyType":11,"taxonomyTypeDesc":12},"01JBSEEDCAT000000000000010","后端与接口",[253,254,255,256,257,125,258,259,260,261,262,226,263,264,265,266,267,268,232,269,235],"llms","vLLM","serving llms vllm","serving","memory","openai api","FP8","Inference Serving","Transformers","api","latency","AWQ","GPTQ","throughput","with","openai-compatible","batching",{"skillNo":271,"name":245,"nameZh":18,"description":246,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":272,"categories":273,"keywords":275,"publishedAt":236},"01M6KQ38RR2WHG23XVQFDW7CQS","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fvllm\u002FSKILL.md",[274],{"categoryNo":250,"categoryName":251,"taxonomyType":11,"taxonomyTypeDesc":12},[253,254,255,256,257,125,258,259,260,261,262,226,263,264,265,266,267,268,232,269,235],{"skillNo":277,"name":278,"nameZh":18,"description":279,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":280,"categories":281,"keywords":282,"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",[],[283,284,285,286,287,259,288,260,289,226,290,263,291,292,293,266,267,294,232,233,269,235],"tensorrt llm","tensorrt","pytorch","deployment","Optimization","NVIDIA","llm inference","scaling","TensorRT-LLM","multi","H100","you",{"skillNo":296,"name":278,"nameZh":18,"description":279,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":297,"categories":298,"keywords":299,"publishedAt":236},"01M6KQ30X8607AHGHWM5EDMV9A","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Ftensorrt-llm\u002FSKILL.md",[],[283,284,285,286,287,259,288,260,289,226,290,263,291,292,293,266,267,294,232,233,269,235],{"skillNo":301,"name":302,"nameZh":18,"description":303,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":304,"categories":305,"keywords":306,"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",[],[307,308,261,309,310,311,312,313,314,315,263,316,317,318,319,320,321,322,323,324,235],"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":326,"name":302,"nameZh":18,"description":303,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":327,"categories":328,"keywords":329,"publishedAt":236},"01M6KQ2NXC3KJS8MJ51CY9HCY3","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fspeculative-decoding\u002FSKILL.md",[],[307,308,261,309,310,311,312,313,314,315,263,316,317,318,319,320,321,322,323,324,235],{"skillNo":331,"name":332,"nameZh":18,"description":333,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":334,"categories":335,"keywords":336,"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",[],[337,338,253,339,340,288,260,341,342,343,344,345,308,346,261,347,348,349,294,350,235],"SGLang","agentic","workflows","agentic workflows","Structured Generation","RadixAttention","Prefix Caching","Constrained Decoding","JSON Output","Production Scale","xai","caching","regex","at",{"skillNo":352,"name":332,"nameZh":18,"description":333,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":353,"categories":354,"keywords":355,"publishedAt":236},"01M6KQ2CQNSSM76QDB2MT9NQCN","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fsglang\u002FSKILL.md",[],[337,338,253,339,340,288,260,341,342,343,344,345,308,346,261,347,348,349,294,350,235],{"skillNo":357,"name":358,"nameZh":18,"description":359,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":360,"categories":361,"keywords":362,"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",[],[224,363,308,261,364,316,365,317,366,367,368,369,370,371,372,373,374,232,375,376,235],"Model Pruning","methods","Model Compression","Wanda","SparseGPT","Sparsity","N:M Sparsity","One-Shot Pruning","Structured Pruning","Unstructured Pruning","unstructured","techniques","reduce","accuracy",{"skillNo":378,"name":358,"nameZh":18,"description":359,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":379,"categories":380,"keywords":381,"publishedAt":236},"01M6KQ23BGG9JP587B4PMHBYMD","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fmodel-pruning\u002FSKILL.md",[],[224,363,308,261,364,316,365,317,366,367,368,369,370,371,372,373,374,232,375,376,235],{"skillNo":383,"name":384,"nameZh":18,"description":385,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":386,"categories":387,"keywords":388,"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",[],[389,390,391,257,285,288,260,392,393,394,395,396,397,398,399,400,401,402,232,233,235],"cpp","Llama","llama cpp","apple","bit","silicon","CPU","GGUF","llama.cpp","CPU Inference","Edge Deployment","Non-NVIDIA","AMD GPUs","Intel GPUs",{"skillNo":404,"name":384,"nameZh":18,"description":385,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":405,"categories":406,"keywords":407,"publishedAt":236},"01M6KQ1V8C4MCT1HDM2JNPER05","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fllama-cpp\u002FSKILL.md",[],[389,390,391,257,285,288,260,392,393,394,395,396,397,398,399,400,401,402,232,233,235],{"skillNo":409,"name":410,"nameZh":18,"description":411,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":412,"categories":413,"keywords":414,"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",[],[415,416,417,253,257,339,287,418,254,261,224,226,419,393,420,365,267,232,233,421,235],"Quantization","HQQ","hqq quantization","data","efficiency","Memory Efficiency","calibration",{"skillNo":423,"name":410,"nameZh":18,"description":411,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":424,"categories":425,"keywords":426,"publishedAt":236},"01M6KQ1J6AJWBQ2TAVZZWS4HFJ","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fhqq\u002FSKILL.md",[],[415,416,417,253,257,339,287,418,254,261,224,226,419,393,420,365,267,232,233,421,235],{"skillNo":428,"name":429,"nameZh":18,"description":430,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":431,"categories":432,"keywords":433,"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",[],[265,253,257,434,287,435,308,436,261,437,438,393,439,440,441,442,294,232,233,376,235],"training","FP16","perplexity","tuning","auto","4-Bit","Memory Optimization","Consumer GPUs","Group-Wise Quantization",{"skillNo":444,"name":429,"nameZh":18,"description":430,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":445,"categories":446,"keywords":447,"publishedAt":236},"01M6KQ17X75YY66NBWNWFD6KEE","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Inference-and-Optimization\u002Fgptq\u002FSKILL.md",[],[265,253,257,434,287,435,308,436,261,437,438,393,439,440,441,442,294,232,233,376,235],10879,18]