[{"data":1,"prerenderedAt":386},["ShallowReactive",2],{"project-19025709":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":28,"sentiment":29,"stance_mix":30,"scored_comments":35,"aspects":36,"strip":46,"post_text":47,"what":48,"said":110,"maker_in_thread":246,"shape":257,"grid":260,"attempts":346,"same_week":347,"maker_shows":38,"maker_hidden":378,"reader_version":379,"card":380},"19025709","AiFiddle","Deep Learning GUI to Create, Train and Visualize Models in a Browser","title","Show HN: Deep Learning GUI to Create, Train and Visualize Models in a Browser","https:\u002F\u002Fbeta.aifiddle.io\u002F","beta.aifiddle.io","0mbre",1548764369,2019,216,44,178,2195,9,"notable","ai","deep learning gui","web_app",[24,25,26,27],"deep learning","neural network","visual model builder","keras export","mixed",0.178,{"positive":31,"neutral":32,"mixed":33,"critical":34},10,8,5,4,27,{"design":37,"execution":40,"idea":42,"performance":44,"docs":45},{"positive":38,"critical":39},1,0,{"positive":39,"critical":41},3,{"positive":43,"critical":39},6,{"positive":39,"critical":38},{"positive":39,"critical":39},[41,41,39,33,38,33,41,39,41,41,39,41,39,38,39,38,39,41,39,33,33,41,33,39,41,39,33,41,41,41,33,33,39,41,39,39,39,33,38,33],null,{"problem":49,"audience":65,"pricing":47,"motivation":76,"time_to_build":47,"stack":81,"openness":108,"stage":109,"team_size":108},{"value":50,"source":51,"evidence":52},"Building and debugging neural networks visually, especially tensor shape mismatches","commenter",[53,57,61],{"id":54,"by":55,"text":56},"19035934","WhiteNoiz3","Cool, project - I really believe that we need more visual tools for building and debugging neural nets. I had a hard time figuring out how to delete a layer (the icons on the…",{"id":58,"by":59,"text":60},"19033648","ovi256","If by flow you mean the tensor sizes in different layers, that's a huge help for debugging, or when rewriting the net in a another framework. For debugging, you're otherwise…",{"id":62,"by":63,"text":64},"19025723","k4renio","Very nice concept. This has the potential to be a great teach \u002F prototyping tool",{"value":66,"source":51,"evidence":67},"Students, teachers and people learning machine learning",[68,72,73],{"id":69,"by":70,"text":71},"19033209","vinayan3","Really awesome and will be super helpful as a teaching aide for ML\u002FDeep learning classes!",{"id":62,"by":63,"text":64},{"id":74,"by":11,"text":75},"19034422","1- Indeed. Part of the vision for AiFiddle is to make a educational tool. 2- Yep fairly easy win too as I am using Cognito for auth. Thanks for the feedback",{"value":77,"source":78,"evidence":79},"Make an educational tool for learning neural networks visually","maker",[80],{"id":74,"by":11,"text":75},[82,91,96,104],{"value":83,"source":78,"evidence":84},"React",[85,88],{"id":86,"by":11,"text":87},"19035755","Thanks! The frontend is made with React and Typescript, backend is AWS serverless",{"id":89,"by":11,"text":90},"19034413","Thanks. GUI is built using React with Typescript. Backend is serverless.",{"value":92,"source":78,"evidence":93},"TypeScript",[94,95],{"id":86,"by":11,"text":87},{"id":89,"by":11,"text":90},{"value":97,"source":78,"evidence":98},"AWS serverless (Lambda)",[99,100],{"id":86,"by":11,"text":87},{"id":101,"by":102,"text":103},"19034253","m4nu","AWS lambda",{"value":105,"source":78,"evidence":106},"AWS Cognito for auth",[107],{"id":74,"by":11,"text":75},"unknown","beta",{"praise":111,"objections":143,"requests":176,"questions":220,"alternatives":236},[112,121,136],{"value":113,"source":51,"evidence":114},"Useful as a teaching and prototyping tool",[115,116,117],{"id":69,"by":70,"text":71},{"id":62,"by":63,"text":64},{"id":118,"by":119,"text":120},"19037142","edshiro","Super awesome! I had a similar idea about a year ago but never got to implement it. We need more visual and straightforward ways to create neural network architectures. Can you…",{"value":122,"source":51,"evidence":123},"Slick, well-done interface",[124,128,132],{"id":125,"by":126,"text":127},"19035654","samzer","This is so slick!",{"id":129,"by":130,"text":131},"19033202","amitport","This is really well done. I've got to mention two missing features that I find important: 1 - help popups: it will be nice if terms had a small help icon that can display a…",{"id":133,"by":134,"text":135},"19033157","Rainymood","What is the tech stack used for the website\u002Fapp? Looks great btw.",{"value":137,"source":51,"evidence":138},"Drawing digits and seeing recognition is fun",[139],{"id":140,"by":141,"text":142},"19034243","hacker_9","Wow this is great! Makes the process so much more interactive, being able to draw your own digits and it gets recognised is great. Some small issues I ran into for your todo…",[144,148,162,169],{"value":145,"source":51,"evidence":146},"Layer icons are tiny and deleting layers is unclear",[147],{"id":54,"by":55,"text":56},{"value":149,"source":51,"evidence":150},"Training errors and bugs reported",[151,155,159],{"id":152,"by":153,"text":154},"19034234","mattlondon","Really nice - really like it for just fiddling around with models. One missing feature would be an \"undo\" function - I accidentally deleted a layer and couldn't find a way to…",{"id":156,"by":157,"text":158},"19035318","kaybe","Yes please! I have no idea what I'm doing (ok only a little), and finally got the dimensions to match and have trainable layers, and now I am informed that 't.slice is not a…",{"id":160,"by":157,"text":161},"19035607","As I said, I don't really know what I'm doing.. The editor crashed while I was playing with it, so I remade the model here: (still the same error)…",{"value":163,"source":51,"evidence":164},"No mobile support",[165],{"id":166,"by":167,"text":168},"19035232","Nashooo","Hi, the message on mobile has two errors: \"Appologies AiFiddle currently does not support mobile device. Do you want to get notified on futur updates ?\" 'Apologies' and 'future'.…",{"value":170,"source":51,"evidence":171},"Shape changes do not cascade through the graph",[172],{"id":173,"by":174,"text":175},"19036319","glastra","Editing a layer and changing the number of filters, stride or kernel size effectively changes the shape of that layer, but it should cascade a re-calculation down the path, which…",[177,191,202,212,216],{"value":178,"source":51,"evidence":179},"Export to Keras, PyTorch or TensorFlow code",[180,184,188],{"id":181,"by":182,"text":183},"19033497","bicubic","Awesome work. Now make it emit keras\u002Ftf\u002Ftorch code!",{"id":185,"by":186,"text":187},"19033722","ivan_ah","Had the same idea; opened a \"feature request\" here: https:\u002F\u002Fgithub.com\u002Fm4nuC\u002Faifiddle-issues\u002Fissues\u002F6 Not sure how easy it will be to support all the types of layers on all…",{"id":189,"by":11,"text":190},"19034528","Thanks for taking the time. Keras functional API export coming soon. The rest will follow.",{"value":192,"source":51,"evidence":193},"Custom dataset support",[194,198,199],{"id":195,"by":196,"text":197},"19038273","abrichr","The demo video is very compelling! Do you support proprietary datasets? A small suggestion: the audio on the demo video was poor quality and difficult to understand. Perhaps a…",{"id":118,"by":119,"text":120},{"id":200,"by":11,"text":201},"19043409","Thanks! Custom dataset indeed on the todo list. The feature is half done already, I should be able to release it soon.",{"value":203,"source":51,"evidence":204},"Export to Colab",[205,209],{"id":206,"by":207,"text":208},"19035672","syntaxing","This is super awesome! Being able to export this to colab would be amazing. I'm really looking forward to this feature! Are you willing to share which web framework you used to…",{"id":210,"by":102,"text":211},"19034164","In the backlog. With export to Colab.",{"value":213,"source":51,"evidence":214},"Undo for deleted layers",[215],{"id":152,"by":153,"text":154},{"value":217,"source":51,"evidence":218},"Automatic reshape layers or suggestions",[219],{"id":54,"by":55,"text":56},[221,229],{"value":222,"source":51,"evidence":223},"Is the tech stack used?",[224,225],{"id":133,"by":134,"text":135},{"id":226,"by":227,"text":228},"19033869","pplonski86","I'm also interested in tech stack used. What are you using in the backend?",{"value":230,"source":51,"evidence":231},"Why visualizing the tensor flow matters",[232],{"id":233,"by":234,"text":235},"19033593","ztratar","I'm not a machine learning engineer, so I'll ask the dumb question: why is visualizing this flow important? What's the benefit?",[237],{"value":238,"source":51,"evidence":239},"Google Colab",[240,243],{"id":241,"by":182,"text":242},"19035242","Anyone who is messing around with convolutional networks is almost guaranteed to have anaconda installed. The convenience of a few less clicks in exchange for platform lock-in is…",{"id":244,"by":182,"text":245},"19034319","I'd encourage everyone not to support colab or any other proprietary lock-in platforms. But do as you will.",{"comments":247,"answered":248,"key":249},14,7,[250,253,256],{"id":251,"text":252},"19034408","Good catch. Thanks. Indeed due for a perf update.",{"id":254,"text":255},"19034428","Thanks for reporting. I haven't done a lot of testing outside of Chrome, will check this out.",{"id":74,"text":75},{"commenters":258,"max_depth":43,"maker_replies":247,"first_comment_mins":41,"lifespan_hours":259},23,46,[261,263,265,267,270,272,273,275,277,278,279,280,282,283,285,286,288,292,295,298,301,302,304,307,309,311,313,315,317,319,322,323,324,325,327,328,329,330,332,334,335,339,341,343],[54,39,41,262,55,39],"project\u002Fmixed\u002Fdesign",[140,38,41,264,141,39],"project\u002Fmixed\u002Fexecution",[251,38,39,266,11,38],"maker\u002Fneutral\u002Fexecution",[69,268,33,269,70,39],2,"project\u002Fpositive\u002Fidea",[166,41,38,271,167,39],"project\u002Fcritical\u002Fexecution",[62,34,33,269,63,39],[233,33,41,274,234,39],"project\u002Fneutral\u002Fidea",[58,33,39,276,59,39],"commenter\u002Fneutral\u002Fidea",[152,43,41,264,153,39],[254,43,39,266,11,38],[129,248,41,262,130,39],[74,248,39,281,11,38],"maker\u002Fneutral\u002Fdesign",[156,248,41,264,157,39],[284,248,39,266,11,38],"19035402",[160,248,38,271,157,39],[287,248,39,266,11,38],"19035663",[289,32,38,290,291,39],"19033828","project\u002Fcritical\u002Fperformance","hnuser123456",[293,32,39,294,11,38],"19035989","maker\u002Fneutral\u002Fperformance",[296,32,41,297,291,39],"19036599","project\u002Fneutral\u002Fperformance",[299,32,39,300,11,38],"19043311","maker\u002Fpositive\u002Fperformance",[118,18,33,269,119,39],[200,18,39,303,11,38],"maker\u002Fpositive\u002Fidea",[305,18,33,306,119,39],"19043869","maker\u002Fpositive\u002Fnone",[195,31,41,308,196,39],"project\u002Fneutral\u002Fdocs",[206,310,33,269,207,39],11,[86,310,39,312,11,38],"maker\u002Fneutral\u002Fnone",[314,310,41,312,207,39],"19035917",[316,310,39,312,11,38],"19036296",[318,310,33,306,207,39],"19036799",[133,320,41,321,134,39],12,"project\u002Fneutral\u002Fnone",[226,320,41,321,227,39],[101,320,41,321,102,39],[89,320,39,312,11,38],[181,326,33,269,182,39],13,[185,326,33,269,186,39],[189,326,39,303,11,38],[210,326,41,274,102,39],[244,326,39,331,182,39],"commenter\u002Fcritical\u002Fidea",[333,326,39,276,11,38],"19034914",[241,326,39,331,182,39],[336,247,33,337,338,39],"19025743","project\u002Fpositive\u002Fnone","dadel",[173,340,38,271,174,39],15,[342,340,39,266,11,38],"19036706",[125,344,33,345,126,39],16,"project\u002Fpositive\u002Fdesign",[],[348,353,358,363,368,373],{"id":349,"name":350,"points":351,"category":352},"19025560","Open Startups (postmake.io\u002Fopen)",475,"data",{"id":354,"name":355,"points":356,"category":357},"19017270","X-spreadsheet",347,"languages",{"id":359,"name":360,"points":361,"category":362},"19023196","Apprise",163,"infrastructure",{"id":364,"name":365,"points":366,"category":367},"18997160","MadCO2",146,"hardware",{"id":369,"name":370,"points":371,"category":372},"19013445","Notepin",144,"productivity",{"id":374,"name":375,"points":376,"category":377},"18995551","Queueing theory",126,"learning_science",false,"v2",{"id":381,"name":5,"line":6,"ai":378,"cat":20,"year":13,"points":14,"comments":15,"rec":28,"s":382,"top":18,"kw":383,"maker":11,"domain":10,"tier":19,"attempts":39,"strip":384,"url":9,"mix":385,"praise":113,"objection":145,"sub":21},19025709,0.18,[24,25,26,27],"3305153033030101030553503053335503000515",[31,32,33,34],1794036574397]