[{"data":1,"prerenderedAt":199},["ShallowReactive",2],{"project-49354664":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":10,"maker":11,"posted_at":12,"year":13,"points":14,"comments":15,"rank_in_year":16,"shown_in_year":17,"top":18,"tier":19,"category":20,"subcategory":21,"format":22,"keywords":23,"reception":27,"sentiment":28,"stance_mix":29,"scored_comments":33,"aspects":34,"strip":40,"post_text":42,"what":43,"said":69,"maker_in_thread":123,"shape":129,"grid":133,"attempts":164,"same_week":165,"maker_shows":36,"maker_hidden":192,"reader_version":193,"card":194},"49354664","modelmap","Interactive, animated architecture of any HuggingFace models","title","Show HN: Interactive, animated architecture of any HuggingFace models","https:\u002F\u002Fmodelmap.cc","modelmap.cc","lizhaoliu",1787097456,2026,113,12,599,4366,14,"notable","ai","model visualization","web_app",[24,25,21,26],"huggingface models","neural network architecture","kv cache estimator","mixed",0.25,{"positive":30,"neutral":31,"mixed":32,"critical":31},5,2,0,9,{"design":35,"execution":37,"performance":38,"idea":39},{"positive":32,"critical":36},1,{"positive":32,"critical":36},{"positive":36,"critical":32},{"positive":36,"critical":32},[41,32,41,32,30,30,30,32,30,36,36,30],3,null,{"problem":44,"audience":42,"pricing":42,"motivation":42,"time_to_build":42,"stack":56,"openness":68,"stage":68,"team_size":68},{"value":45,"source":46,"evidence":47},"Hard to see how a HuggingFace model is architected and what it costs to serve","commenter",[48,52],{"id":49,"by":50,"text":51},"49355324","alansml","Love it, I thought it only shows the \"guts\" of a model, turns out it also estimates the cost of serving that model.",{"id":53,"by":54,"text":55},"49355861","Hacker_Yogi","Visually inspecting how a particular model is architected is quite interesting. Thank you for coming up with this!",[57,64],{"value":58,"source":59,"evidence":60},"PyTorch","maker",[61],{"id":62,"by":11,"text":63},"49355818","Hi! Not from the config alone, the config just tells transformers which classes to build. I let it actually build the model, but on PyTorch's meta device, so every module and…",{"value":65,"source":59,"evidence":66},"HuggingFace transformers",[67],{"id":62,"by":11,"text":63},"unknown",{"praise":70,"objections":80,"requests":95,"questions":103,"alternatives":115},[71],{"value":72,"source":46,"evidence":73},"Interesting way to inspect model architecture; cost estimation is a useful addition",[74,75,76],{"id":49,"by":50,"text":51},{"id":53,"by":54,"text":55},{"id":77,"by":78,"text":79},"49355199","mojosmojo","extremely cool. I’ve had Claude walk model architectures to debug Loras and fine tunes, but this is delightful",[81,88],{"value":82,"source":46,"evidence":83},"Search appears broken",[84],{"id":85,"by":86,"text":87},"49360736","virajk_31","I think search is broken..",{"value":89,"source":46,"evidence":90},"Not as interactive or animated as the description suggested",[91],{"id":92,"by":93,"text":94},"49364221","sam1234apter","when you mean interactive and animated i had high hopes",[96],{"value":97,"source":46,"evidence":98},"Surface tensor shapes and layer-level parameter counts",[99],{"id":100,"by":101,"text":102},"49355460","xms17189","How are you generating the architecture graph from model configs, and do you plan to surface tensor shapes or layer-level parameter counts as well?",[104,108],{"value":105,"source":46,"evidence":106},"How the graph is generated from model configs",[107],{"id":100,"by":101,"text":102},{"value":109,"source":46,"evidence":110},"Whether specific layers can be fine-tuned with LoRA",[111],{"id":112,"by":113,"text":114},"49359257","dotancohen","I'd love to hear more about that. Are you able to fine tune specific layers? Specifically, one particular otherwise excellent model I use has an alignment problem (sycophancy)…",[116],{"value":117,"source":46,"evidence":118},"hfviewer.com",[119],{"id":120,"by":121,"text":122},"49356427","srvmshr","I saw a similar page a month back. The design aesthetics are bit different https:\u002F\u002Fhfviewer.com\u002F",{"comments":31,"answered":36,"key":124},[125,128],{"id":126,"text":127},"49356831","Oh nice, hadn't seen it. It looks good, and they've clearly got some decent stuff I don't (chrome extension, embedding in model cards, family pages). The main difference appears…",{"id":62,"text":63},{"commenters":130,"max_depth":31,"maker_replies":31,"first_comment_mins":131,"lifespan_hours":132},11,63,24,[134,136,138,140,141,143,145,148,150,154,157,160],[120,32,41,135,121,32],"project\u002Fneutral\u002Fdesign",[126,32,32,137,11,36],"other\u002Fneutral\u002Fexecution",[100,36,41,139,101,32],"project\u002Fneutral\u002Fexecution",[62,36,32,139,11,36],[49,31,30,142,50,32],"project\u002Fpositive\u002Fperformance",[53,41,30,144,54,32],"maker\u002Fpositive\u002Fidea",[77,146,30,147,78,32],4,"project\u002Fpositive\u002Fnone",[112,146,32,149,113,32],"other\u002Fneutral\u002Fnone",[151,30,30,152,153,32],"49355253","maker\u002Fpositive\u002Fnone","bicepjai",[85,155,36,156,86,32],6,"project\u002Fcritical\u002Fexecution",[92,158,36,159,93,32],7,"project\u002Fcritical\u002Fdesign",[161,162,30,147,163,32],"49368621",8,"padovan93",[],[166,171,176,180,184,188],{"id":167,"name":168,"points":169,"category":170},"49327643","Desktopcolors.com",166,"learning_science",{"id":172,"name":173,"points":174,"category":175},"49355607","Lucasartsifier",161,"dev_tools",{"id":177,"name":178,"points":179,"category":20},"49307700","ThoughtDAG",136,{"id":181,"name":182,"points":183,"category":175},"49364296","terminal-code",108,{"id":185,"name":186,"points":187,"category":175},"49301215","LuaCAD",106,{"id":189,"name":190,"points":191,"category":20},"49303046","Mole",100,false,"v2",{"id":195,"name":5,"line":6,"ai":192,"cat":20,"year":13,"points":14,"comments":15,"rec":27,"s":28,"top":18,"kw":196,"maker":11,"domain":10,"tier":19,"attempts":32,"strip":197,"url":9,"mix":198,"praise":72,"objection":82,"sub":21},49354664,[24,25,21,26],"303055505115",[30,31,32,31],1794036708234]