[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skills:stats":3,"skills:sce6":8,"skills:pro6":55,"skills:creators9":100,"skills:all:stars:26759:":195},{"skillCount":4,"professionCount":5,"scenarioCount":6,"creatorCount":7},482556,11,7,5953,[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",1,39675,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",26798,[],{"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",13847,[],{"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",6403,[],{"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",6063,[],{"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",5574,[],[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",13341,[],{"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",7902,[],{"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",5005,[],{"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",3978,[],{"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",3454,[],{"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",3337,[],{"records":101,"total":7,"current":16,"size":194},[102,114,123,132,142,152,162,172,182],{"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",4300,236,832929,"2026-11-15 04:18:35",{"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.",369,32,437811,"2026-11-15 03:04:02",{"accountNo":124,"login":125,"displayName":126,"avatarUrl":127,"accountType":107,"accountTypeDesc":108,"profileDescription":18,"skillCount":128,"repoCount":129,"professionCount":5,"scenarioCount":6,"totalStars":130,"updatedAt":131},"01M697XGDPXJAVQ8391V2BCYTE","anthropics","Anthropic","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F76263028?v=4",716,18,427323,"2026-11-15 05:52:49",{"accountNo":133,"login":134,"displayName":135,"avatarUrl":136,"accountType":107,"accountTypeDesc":108,"profileDescription":18,"skillCount":137,"repoCount":138,"professionCount":6,"scenarioCount":139,"totalStars":140,"updatedAt":141},"01M6AC29YGCR4S2NTJFV0G8B81","openai","OpenAI","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F14957082?v=4",177,23,6,344400,"2026-11-15 03:07:28",{"accountNo":143,"login":144,"displayName":145,"avatarUrl":146,"accountType":107,"accountTypeDesc":108,"profileDescription":18,"skillCount":147,"repoCount":148,"professionCount":149,"scenarioCount":6,"totalStars":150,"updatedAt":151},"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":153,"login":154,"displayName":155,"avatarUrl":156,"accountType":107,"accountTypeDesc":108,"profileDescription":157,"skillCount":158,"repoCount":159,"professionCount":5,"scenarioCount":6,"totalStars":160,"updatedAt":161},"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":163,"login":164,"displayName":165,"avatarUrl":166,"accountType":107,"accountTypeDesc":108,"profileDescription":167,"skillCount":168,"repoCount":169,"professionCount":5,"scenarioCount":6,"totalStars":170,"updatedAt":171},"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":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",87,2,10,188029,"2026-11-15 06:21:09",{"accountNo":183,"login":184,"displayName":185,"avatarUrl":186,"accountType":187,"accountTypeDesc":188,"profileDescription":189,"skillCount":190,"repoCount":191,"professionCount":139,"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. ",340,3,175019,"2026-11-15 03:57:03",9,{"records":196,"total":4,"current":480,"size":129},[197,232,261,268,292,297,321,326,352,357,379,384,401,406,429,435,452,457],{"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":210,"publishedAt":231},"01M6KQN0YZPJ8SRTNQX4PCFP9Z","openrlhf-training","High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.","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-Training\u002Fopenrlhf\u002FSKILL.md",[],[211,212,213,214,215,216,217,218,219,220,221,222,223,224,225,226,227,228,229,230],"training","OpenRLHF","openrlhf training","performance","architecture","framework","DeepSpeed","Distributed Training","ZeRO","Post-Training","RLHF","PPO","RLOO","Ray","vLLM","Transformers","use","high-performance","orchestra","for","2026-11-10 10:12:30",{"skillNo":233,"name":234,"nameZh":18,"description":235,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":236,"categories":237,"keywords":242,"publishedAt":231},"01M6KQMVW9KHRK5BNBH9XX5MJ6","nemo-curator","GPU-accelerated data curation for LLM training. Supports text\u002Fimage\u002Fvideo\u002Faudio. Features fuzzy deduplication (16× faster), quality filtering (30+ heuristics), semantic deduplication, PII redaction, NSFW detection. Scales across GPUs with RAPIDS. Use for preparing high-quality training datasets, cleaning web data, or deduplicating large corpora.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fnemo-curator\u002Fnemo-curator\u002FSKILL.md",[238,241],{"categoryNo":239,"categoryName":240,"taxonomyType":11,"taxonomyTypeDesc":12},"01JBSEEDCAT000000000000016","图像与多媒体",{"categoryNo":73,"categoryName":74,"taxonomyType":58,"taxonomyTypeDesc":59},[243,244,245,211,246,247,248,249,250,251,252,253,254,255,256,257,258,259,229,260],"NeMo","nemo curator","image","multimodal","NVIDIA","data","deduplication","heuristics","pii","dask","Video","Audio","text","across","processing","GPU Acceleration","curation","features",{"skillNo":262,"name":234,"nameZh":18,"description":235,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":263,"categories":264,"keywords":267,"publishedAt":231},"01M6KQMPN86GRJBJZC2D3KNK0D","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fnemo-curator\u002FSKILL.md",[265,266],{"categoryNo":239,"categoryName":240,"taxonomyType":11,"taxonomyTypeDesc":12},{"categoryNo":73,"categoryName":74,"taxonomyType":58,"taxonomyTypeDesc":59},[243,244,245,211,246,247,248,249,250,251,252,253,254,255,256,257,258,259,229,260],{"skillNo":269,"name":270,"nameZh":18,"description":271,"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":274,"publishedAt":231},"01M6KQMHBB6SBRB7NQ71E4H7TG","moe-training","Train Mixture of Experts (MoE) models using DeepSpeed or HuggingFace. Use when training large-scale models with limited compute (5× cost reduction vs dense models), implementing sparse architectures like Mixtral 8x7B or DeepSeek-V3, or scaling model capacity without proportional compute increase. Covers MoE architectures, routing mechanisms, load balancing, expert parallelism, and inference optimization.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fmoe-training\u002Fmoe-training\u002FSKILL.md",[],[211,275,276,217,277,226,278,279,280,281,282,283,284,285,286,287,288,289,229,290,291],"moe","moe training","Optimization","model","expert","scaling","architectures","Accelerate","Emerging Techniques","Mixture Of Experts","Sparse Models","Expert Parallelism","Mixtral","Efficient Training","experts","implementing","when",{"skillNo":293,"name":270,"nameZh":18,"description":271,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":294,"categories":295,"keywords":296,"publishedAt":231},"01M6KQMAMXND3YWMDB0ERJM0T2","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fmoe-training\u002FSKILL.md",[],[211,275,276,217,277,226,278,279,280,281,282,283,284,285,286,287,288,289,229,290,291],{"skillNo":298,"name":299,"nameZh":18,"description":300,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":301,"categories":302,"keywords":303,"publishedAt":231},"01M6KQM4FBJE72KSFS7Q4V480Q","model-merging","Merge multiple fine-tuned models using mergekit to combine capabilities without retraining. Use when creating specialized models by blending domain-specific expertise (math + coding + chat), improving performance beyond single models, or experimenting rapidly with model variants. Covers SLERP, TIES-Merging, DARE, Task Arithmetic, linear merging, and production deployment strategies.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fmodel-merging\u002Fmodel-merging\u002FSKILL.md",[],[278,304,214,305,306,226,307,308,309,310,283,311,312,313,314,315,316,317,318,319,320,291],"Model Merging","merge","capabilities","math","specialized","strategies","fusion","Mergekit","SLERP","TIES","DARE","Task Arithmetic","Model Fusion","No Retraining","Multi-Capability","expertise","variants",{"skillNo":322,"name":299,"nameZh":18,"description":300,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":323,"categories":324,"keywords":325,"publishedAt":231},"01M6KQKWEXEDSM6VF6P6SY2QC2","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fmodel-merging\u002FSKILL.md",[],[278,304,214,305,306,226,307,308,309,310,283,311,312,313,314,315,316,317,318,319,320,291],{"skillNo":327,"name":328,"nameZh":18,"description":329,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":330,"categories":331,"keywords":332,"publishedAt":231},"01M6KQKPB3Z5XJJJDXMQRAGXZM","ml-training-recipes","Battle-tested PyTorch training recipes for all domains — LLMs, vision, diffusion, medical imaging, protein\u002Fdrug discovery, spatial omics, genomics. Covers training loops, optimizer selection (AdamW, Muon), LR scheduling, mixed precision, debugging, and systematic experimentation. Use when training or fine-tuning neural networks, debugging loss spikes or OOM, choosing architectures, or optimizing GPU throughput.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fml-training-recipes\u002Fml-training-recipes\u002FSKILL.md",[],[333,211,334,335,336,337,338,339,340,341,342,343,344,345,346,347,348,349,350,351,291],"ml","recipes","ml training recipes","llms","neural networks","pytorch","debugging","tuning","selection","optimizer","medical","systematic","domains","experimentation","omics","imaging","loops","Muon","AdamW",{"skillNo":353,"name":328,"nameZh":18,"description":329,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":354,"categories":355,"keywords":356,"publishedAt":231},"01M6KQKFN8JM39GVHFZ97EJXSW","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fml-training-recipes\u002FSKILL.md",[],[333,211,334,335,336,337,338,339,340,341,342,343,344,345,346,347,348,349,350,351,291],{"skillNo":358,"name":359,"nameZh":18,"description":360,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":361,"categories":362,"keywords":363,"publishedAt":231},"01M6KQK8F2BMD1FTGP9Z6YXK07","miles-rl-training","Provides guidance for enterprise-grade RL training using miles, a production-ready fork of slime. Use when training large MoE models with FP8\u002FINT4, needing train-inference alignment, or requiring speculative RL for maximum throughput.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fmiles\u002Fmiles\u002FSKILL.md",[],[211,364,365,366,367,368,369,226,370,227,371,372,373,374,375,376,377,378,229,230,291],"rl","miles rl training","miles","FP8","Enterprise","SGLang","learning","router","ready","alignment","slime","megatron","Megatron-LM","throughput","production-ready",{"skillNo":380,"name":359,"nameZh":18,"description":360,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":381,"categories":382,"keywords":383,"publishedAt":231},"01M6KQK36YJNBJ379SE7T9GBN6","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fmiles\u002FSKILL.md",[],[211,364,365,366,367,368,369,226,370,227,371,372,373,374,375,376,377,378,229,230,291],{"skillNo":385,"name":386,"nameZh":18,"description":387,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":388,"categories":389,"keywords":390,"publishedAt":231},"01M6KQJXY9HPMWRB7PXBEF29BC","training-llms-megatron","Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B parameters, need maximum GPU efficiency (47% MFU on H100), or require tensor\u002Fpipeline\u002Fsequence\u002Fcontext\u002Fexpert parallelism. Production-ready framework used for Nemotron, LLaMA, DeepSeek.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fmegatron-core\u002Fmegatron-core\u002FSKILL.md",[],[336,211,375,391,216,392,278,393,372,279,394,309,395,396,397,398,286,399,378,400,291],"training llms megatron","pipeline","context","efficiency","apex","H100","Megatron-Core","Model Parallelism","advanced","sequence",{"skillNo":402,"name":386,"nameZh":18,"description":387,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":403,"categories":404,"keywords":405,"publishedAt":231},"01M6KQJRM8H767ZSZD0CHWK87Y","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fmegatron-core\u002FSKILL.md",[],[336,211,375,391,216,392,278,393,372,279,394,309,395,396,397,398,286,399,378,400,291],{"skillNo":407,"name":408,"nameZh":18,"description":409,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":410,"categories":411,"keywords":413,"publishedAt":231},"01M6KQJMDQ974E2Q475NCSVJNF","llama-factory","Expert guidance for fine-tuning LLMs with LLaMA-Factory - WebUI no-code, 100+ models, 2\u002F3\u002F4\u002F5\u002F6\u002F8-bit QLoRA, multimodal support","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fllama-factory\u002Fllama-factory\u002FSKILL.md",[412],{"categoryNo":57,"categoryName":60,"taxonomyType":58,"taxonomyTypeDesc":59},[414,415,416,336,417,418,419,420,421,422,423,279,424,425,282,426,427,428,229,230],"Llama","factory","LLaMA Factory","research","support","code","Fine-Tuning","LoRA","QLoRA","HuggingFace","webui","bit","8-Bit","No-Code","with",{"skillNo":430,"name":408,"nameZh":18,"description":409,"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":434,"publishedAt":231},"01M6KQJF9X5E6JGNN82MD9X95H","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fllama-factory\u002FSKILL.md",[433],{"categoryNo":57,"categoryName":60,"taxonomyType":58,"taxonomyTypeDesc":59},[414,415,416,336,417,418,419,420,421,422,423,279,424,425,282,426,427,428,229,230],{"skillNo":436,"name":437,"nameZh":18,"description":438,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":439,"categories":440,"keywords":441,"publishedAt":231},"01M6KQJ22RQ1A2X74MZRZV6SFS","implementing-llms-litgpt","Implements and trains LLMs using Lightning AI's LitGPT with 20+ pretrained architectures (Llama, Gemma, Phi, Qwen, Mistral). Use when need clean model implementations, educational understanding of architectures, or production fine-tuning with LoRA\u002FQLoRA. Single-file implementations, no abstraction layers.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Flitgpt\u002Flitgpt\u002FSKILL.md",[],[336,442,443,290,215,420,421,226,340,444,278,227,445,446,281,447,448,449,450,229,451,291],"LitGPT","implementing llms litgpt","phi","single-file","clean","Model Architecture","Lightning AI","LLM Implementation","layers","single",{"skillNo":453,"name":437,"nameZh":18,"description":438,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":454,"categories":455,"keywords":456,"publishedAt":231},"01M6KQHX4TAYGQWVNRYSE6WK0X","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Flitgpt\u002FSKILL.md",[],[336,442,443,290,215,420,421,226,340,444,278,227,445,446,281,447,448,449,450,229,451,291],{"skillNo":458,"name":459,"nameZh":18,"description":460,"descriptionZh":18,"repositoryFullName":201,"ownerLogin":202,"ownerAvatarUrl":203,"ownerAccountNo":204,"ownerDisplayName":18,"repositoryNo":205,"starsCount":206,"licenseSpdxId":207,"licenseFileUrl":18,"sourceHtmlUrl":461,"categories":462,"keywords":466,"publishedAt":231},"01M6KQHR21Q55VAPKC467XKVEB","knowledge-distillation","Compress large language models using knowledge distillation from teacher to student models. Use when deploying smaller models with retained performance, transferring GPT-4 capabilities to open-source models, or reducing inference costs. Covers temperature scaling, soft targets, reverse KLD, logit distillation, and MiniLLM training strategies.","https:\u002F\u002Fgithub.com\u002Fendman100\u002Fskill-router-nexus\u002Fblob\u002Fmain\u002FAI-Training\u002Fknowledge-distillation\u002Fknowledge-distillation\u002FSKILL.md",[463],{"categoryNo":464,"categoryName":465,"taxonomyType":11,"taxonomyTypeDesc":12},"01JBSEEDCAT000000000000021","资料检索与研究",[467,468,469,214,211,306,470,278,309,280,283,471,472,473,474,475,476,477,478,479,291],"knowledge","distillation","Knowledge Distillation","open","Teacher-Student","MiniLLM","Reverse KLD","Soft Targets","Temperature Scaling","Logit Distillation","Model Transfer","student","teacher",26759]