[{"data":1,"prerenderedAt":695},["ShallowReactive",2],{"project-32167704":3},{"id":4,"name":5,"line":5,"line_source":6,"title":7,"url":8,"domain":9,"maker":10,"posted_at":11,"year":12,"points":13,"comments":14,"rank_in_year":15,"shown_in_year":16,"top":17,"tier":18,"category":19,"subcategory":20,"format":21,"keywords":22,"reception":28,"sentiment":29,"stance_mix":30,"scored_comments":35,"aspects":36,"strip":47,"post_text":50,"what":51,"said":80,"maker_in_thread":176,"shape":185,"grid":188,"attempts":657,"same_week":658,"maker_shows":48,"maker_hidden":687,"reader_version":688,"card":689},"32167704","This Food Does Not Exist","title","Show HN: This Food Does Not Exist","https:\u002F\u002Fnyx-ai.github.io\u002Fstylegan2-flax-tpu\u002F","nyx-ai.github.io","MasterScrat",1658332972,2022,248,172,197,2570,8,"notable","ai","image generation","web_app",[23,24,25,26,27],"stylegan2","gan","food image generation","tpu training","synthetic images","curious",0.042,{"positive":31,"neutral":32,"mixed":33,"critical":34},16,38,3,13,70,{"idea":37,"execution":40,"pricing":43,"design":45,"docs":46},{"positive":38,"critical":39},11,7,{"positive":41,"critical":42},1,4,{"positive":41,"critical":44},0,{"positive":41,"critical":44},{"positive":44,"critical":44},[33,41,44,48,33,49,48,33,42,33,49,33,48,41,33,44,33,33,33,49,48,42,48,41,44,33,42,33,33,44,44,44,44,44,44,44,44,48,41,33],5,2,null,{"problem":50,"audience":50,"pricing":50,"motivation":52,"time_to_build":59,"stack":63,"openness":77,"stage":78,"team_size":79},{"value":53,"source":54,"evidence":55},"Explore how to improve and scale StyleGAN training, especially on TPUs","maker",[56],{"id":57,"by":10,"text":58},"32167747","We have trained four StyleGAN2 image generation models and are releasing checkpoints and training code. We are exploring how to improve\u002Fscale up StyleGAN training, particularly…",{"value":60,"source":54,"evidence":61},"Each model trained in less than 10 hours",[62],{"id":57,"by":10,"text":58},[64,68,73],{"value":65,"source":54,"evidence":66},"StyleGAN2",[67],{"id":57,"by":10,"text":58},{"value":69,"source":54,"evidence":70},"Flax",[71],{"id":72},"post",{"value":74,"source":54,"evidence":75},"TPU",[76],{"id":57,"by":10,"text":58},"open_source","launched","unknown",{"praise":81,"objections":105,"requests":121,"questions":140,"alternatives":153},[82,94],{"value":83,"source":84,"evidence":85},"Images look convincingly food-like and appetizing","commenter",[86,90],{"id":87,"by":88,"text":89},"32169055","ComputerCat","Everything looks delish!",{"id":91,"by":92,"text":93},"32169212","n4bz0r","The food looks great! I suppose these models could use some extra training with dishes, though. The plates and glasses look wobbly, which is an instant giveaway. Otherwise, I can…",{"value":95,"source":84,"evidence":96},"Interesting use for food blogs and stock photos",[97,101],{"id":98,"by":99,"text":100},"32171101","georgeburdell","This is the most disturbing “does not exist” yet. A food blog could write itself",{"id":102,"by":103,"text":104},"32169760","beej71","We're looking at the complete collapse of the stock photography market.",[106,117],{"value":107,"source":84,"evidence":108},"Generated images may closely copy training data",[109,113],{"id":110,"by":111,"text":112},"32168850","forgotusername6","I wonder how close the nearest match from the training data is. Was there a cheesecake that looked almost like these generated images?",{"id":114,"by":115,"text":116},"32171181","jdthedisciple","Looks impressive but I can't escape the notion that surely some of the generated images will be very close to the some of the training images? How am I to assess how original the…",{"value":118,"source":84,"evidence":119},"Plates and glasses look wobbly, a giveaway",[120],{"id":91,"by":92,"text":93},[122,129],{"value":123,"source":84,"evidence":124},"Way to generate a fresh image on demand",[125],{"id":126,"by":127,"text":128},"32171091","herpderperator","Is there a way to trigger a fresh image on demand? That's kind of what I expect when I see a does-not-exist site.",{"value":130,"source":84,"evidence":131},"Information on training dataset size and licensing",[132,136],{"id":133,"by":134,"text":135},"32169127","Sebbecking","How big was your training dataset?",{"id":137,"by":138,"text":139},"32168930","waynesonfire","and what was the licensing for the training data that you used?",[141,146],{"value":142,"source":84,"evidence":143},"How original are outputs compared to training data",[144,145],{"id":114,"by":115,"text":116},{"id":110,"by":111,"text":112},{"value":147,"source":84,"evidence":148},"Are there techniques to detect generated images versus photos",[149],{"id":150,"by":151,"text":152},"32170079","wyldfire","Are there any analysis techniques that can easily distinguish between these and real photographs? Do simple things like edge detections or histograms reveal any anomalies?",[154,162,169],{"value":155,"source":84,"evidence":156},"DALL-E and diffusion models",[157,161],{"id":158,"by":159,"text":160},"32170455","xg15","At least with DALL-E you can be sure the food has a name. For a moment I was worried this would produce vaguely food-like images where on closer look you realise you have no idea…",{"id":57,"by":10,"text":58},{"value":163,"source":84,"evidence":164},"Imagen",[165],{"id":166,"by":167,"text":168},"32170265","fxtentacle","I'm honestly surprised that they trained a StyleGAN. Recently, the Imagen architecture has been show to be both easier in structure, easier to train, and even faster to produce…",{"value":170,"source":84,"evidence":171},"thisxdoesnotexist.com collection",[172],{"id":173,"by":174,"text":175},"32171150","tmountain","Aggregate \"does not exist\" website for anyone who's interested. https:\u002F\u002Fthisxdoesnotexist.com\u002F",{"comments":33,"answered":41,"key":177},[178,179,182],{"id":57,"text":58},{"id":180,"text":181},"32167788","We currently train each model independently, ie we first gather a cookie dataset, train a cookie model then restart from scratch for the next one. That's actually something we're…",{"id":183,"text":184},"32240270","We haven't experimented much with Imagen, but our initial conclusions were that: - It's hard to train to a photorealistic quality (we'd be happy to be proven wrong!) - There is…",{"commenters":186,"max_depth":34,"maker_replies":33,"first_comment_mins":49,"lifespan_hours":187},111,159,[189,193,197,200,202,206,210,213,215,218,220,224,226,228,231,235,238,242,245,249,252,255,256,259,261,264,266,268,270,273,276,279,281,285,287,290,291,295,299,301,303,306,309,312,314,318,321,324,327,331,334,337,338,342,345,348,351,353,355,358,362,365,369,372,376,379,381,384,386,389,392,395,397,400,403,406,409,411,414,417,418,422,426,429,432,434,437,438,440,443,446,450,452,457,461,463,467,470,472,476,480,482,485,488,490,492,494,496,499,502,504,507,510,512,514,517,519,521,524,526,529,531,534,537,540,542,545,547,549,551,554,557,560,562,564,566,568,570,572,574,576,578,580,582,584,587,589,591,594,597,599,602,605,607,609,612,615,617,619,621,624,626,629,632,635,638,641,643,645,648,651,654],[190,44,33,191,192,44],"32169795","project\u002Fneutral\u002Fidea","JadoJodo",[194,44,44,195,196,44],"32169975","commenter\u002Fneutral\u002Fidea","bee_rider",[198,44,44,199,192,44],"32170635","commenter\u002Fneutral\u002Fnone",[201,44,44,199,196,44],"32170889",[203,44,41,204,205,44],"32172157","project\u002Fcritical\u002Fexecution","mikkergp",[207,44,41,208,209,44],"32176209","project\u002Fcritical\u002Fidea","Tomis02",[211,44,44,195,212,44],"32176665","krageon",[214,44,44,195,209,44],"32176780",[216,44,44,195,217,44],"32177893","chii",[219,44,44,195,192,44],"32178969",[221,44,44,222,223,44],"32179998","other\u002Fneutral\u002Fnone","YeGoblynQueenne",[225,44,44,222,209,44],"32182273",[227,44,44,222,223,44],"32203971",[229,44,33,191,230,44],"32174274","woojoo666",[232,44,48,233,234,44],"32174388","project\u002Fpositive\u002Fidea","NoPicklez",[236,44,48,233,237,44],"32175746","Smoosh",[239,44,48,240,241,44],"32176330","project\u002Fpositive\u002Fpricing","jjeaff",[243,44,41,204,244,44],"32176599","krater23",[246,44,41,247,248,44],"32211161","project\u002Fcritical\u002Fnone","fancymcpoopoo",[250,44,48,233,251,44],"32176091","staticassertion",[253,44,48,233,254,44],"32178854","XCSme",[110,41,33,191,111,44],[257,41,41,208,258,44],"32169513","layer8",[260,41,33,191,167,44],"32170150",[262,41,44,195,263,44],"32175643","ad404b8a372f2b9",[265,41,44,195,167,44],"32176987",[267,41,44,195,263,44],"32177207",[269,41,44,195,167,44],"32181471",[271,41,48,233,272,44],"32177661","mejutoco",[274,41,44,195,275,44],"32169649","nh23423fefe",[277,41,44,222,278,44],"32171720","yellowapple",[280,41,33,191,205,44],"32172117",[282,41,44,283,284,44],"32175753","other\u002Fneutral\u002Fidea","nmilo",[286,41,44,222,205,44],"32178849",[288,41,33,289,138,44],"32170041","project\u002Fneutral\u002Fexecution",[114,49,33,191,115,44],[292,49,33,293,294,44],"32173251","project\u002Fneutral\u002Fnone","danuker",[296,49,48,297,298,44],"32174783","project\u002Fpositive\u002Fexecution","hwers",[158,33,33,300,159,44],"project\u002Fmixed\u002Fexecution",[302,33,33,289,167,44],"32170534",[304,33,44,199,305,44],"32171395","dalmo3",[307,33,33,293,308,44],"32175077","uptown",[310,33,33,311,223,44],"32180050","project\u002Fmixed\u002Fdesign",[126,42,33,313,127,44],"project\u002Fneutral\u002Fdesign",[315,42,33,316,317,44],"32173244","project\u002Fneutral\u002Fdocs","CSMastermind",[319,48,41,208,320,44],"32169764","notamy",[322,48,44,283,323,44],"32170419","sovok",[57,325,44,326,10,41],6,"maker\u002Fneutral\u002Fexecution",[328,325,33,329,330,44],"32173524","maker\u002Fneutral\u002Fdocs","goldemerald",[332,325,33,326,333,44],"32174243","forgingahead",[335,39,33,289,336,44],"32167739","gus_massa",[180,39,44,326,10,41],[339,39,33,340,341,44],"32169452","maker\u002Fneutral\u002Fnone","TuringNYC",[343,39,44,199,344,44],"32170164","zorgmonkey",[346,39,48,233,347,44],"32168383","dylan604",[349,39,44,222,350,44],"32169064","ch4s3",[352,39,44,222,347,44],"32169315",[354,39,44,222,350,44],"32169437",[356,17,41,208,357,44],"32168259","andrewmcwatters",[359,360,44,222,361,44],"32169758",9,"derbOac",[363,360,44,222,364,44],"32171157","mike_hock",[366,367,44,222,368,44],"32169844",10,"gffrd",[370,38,44,222,371,44],"32173338","Animats",[373,374,33,191,375,44],"32168492",12,"croes",[377,374,44,199,378,44],"32168914","Spivak",[380,374,44,199,350,44],"32169044",[382,374,44,222,383,44],"32169274","ge96",[385,374,44,222,375,44],"32170825",[387,374,33,191,388,44],"32169392","scifibestfi",[390,374,33,191,391,44],"32169675","bergenty",[393,374,44,199,394,44],"32170408","xbar",[396,374,33,191,391,44],"32173797",[398,374,44,222,399,44],"32169401","minimaxir",[401,374,44,222,402,44],"32170416","raphar",[404,374,44,222,405,44],"32171684","Kye",[407,34,44,222,408,44],"32169695","munificent",[150,410,33,289,151,44],14,[412,410,33,289,413,44],"32170489","daveguy",[415,416,44,222,347,44],"32168396",15,[91,31,33,300,92,44],[419,420,33,316,421,44],"32174520",17,"abidlabs",[423,424,33,316,425,44],"32169522",18,"hammycheesy",[427,424,33,316,428,44],"32170885","sireat",[430,424,33,316,431,44],"32177831","lancebeet",[166,433,41,204,167,44],19,[435,433,41,204,436,44],"32171497","gwern",[183,433,44,326,10,41],[98,439,48,233,99,44],20,[441,439,44,222,442,44],"32171142","hbn",[444,439,44,222,445,44],"32177637","WordAWarning",[447,448,48,233,449,44],"32168324",21,"spacemanmatt",[173,451,44,222,174,44],22,[453,454,48,455,456,44],"32171266",23,"project\u002Fpositive\u002Fnone","rkagerer",[458,459,41,208,460,44],"32169069",24,"golergka",[133,462,33,289,134,44],25,[464,465,48,455,466,44],"32168344",26,"WalterSear",[87,468,48,469,88,44],27,"project\u002Fpositive\u002Fdesign",[137,471,33,293,138,44],28,[473,474,41,247,475,44],"32172153",29,"twic",[477,478,33,293,479,44],"32176962",30,"immmmmm",[102,481,44,222,103,44],31,[483,481,44,222,484,44],"32169849","echelon",[486,481,44,222,487,44],"32170605","smaudet",[489,481,44,222,484,44],"32173931",[491,481,44,222,487,44],"32175786",[493,481,44,222,484,44],"32176269",[495,481,44,222,487,44],"32188959",[497,481,44,222,498,44],"32170610","nixpulvis",[500,481,41,208,501,44],"32170664","jeffreygoesto",[503,481,44,222,484,44],"32173956",[505,481,48,233,506,44],"32172427","Victerius",[508,481,48,233,509,44],"32172616","mysterydip",[511,481,33,293,368,44],"32169866",[513,481,44,222,305,44],"32171045",[515,481,33,313,516,44],"32172651","Cockbrand",[518,481,44,222,368,44],"32181470",[520,481,33,293,368,44],"32169880",[522,481,33,293,523,44],"32170435","robotresearcher",[525,481,33,293,368,44],"32170530",[527,481,44,222,528,44],"32172379","DonHopkins",[530,481,33,289,509,44],"32172593",[532,481,33,289,533,44],"32172809","lelandfe",[535,481,44,222,536,44],"32169917","corrral",[538,481,44,199,539,44],"32170314","blueflow",[541,481,44,199,487,44],"32170428",[543,481,44,199,544,44],"32170548","vehementi",[546,481,44,199,536,44],"32170832",[548,481,44,199,544,44],"32172481",[550,481,44,199,536,44],"32173165",[552,481,44,199,553,44],"32183477","nomel",[555,481,44,199,556,44],"32170579","phpnode",[558,481,44,199,559,44],"32170658","mythrwy",[561,481,44,199,556,44],"32170823",[563,481,44,199,536,44],"32170695",[565,481,44,199,553,44],"32183521",[567,481,44,199,559,44],"32170644",[569,481,44,199,536,44],"32170814",[571,481,44,222,559,44],"32170922",[573,481,44,222,536,44],"32171131",[575,481,44,222,559,44],"32171382",[577,481,44,222,278,44],"32171615",[579,481,44,222,509,44],"32172560",[581,481,44,199,278,44],"32175776",[583,481,44,199,487,44],"32175902",[585,481,44,199,586,44],"32170650","unfunco",[588,481,44,199,536,44],"32170818",[590,481,44,199,586,44],"32172460",[592,481,44,222,593,44],"32171042","gadtfly",[595,481,44,199,596,44],"32170454","konschubert",[598,481,44,199,536,44],"32171121",[600,481,44,199,601,44],"32174969","stickfigure",[603,481,44,199,604,44],"32175103","ithkuil",[606,481,44,199,601,44],"32175380",[608,481,44,199,487,44],"32175845",[610,481,44,199,611,44],"32172569","whatshisface",[613,481,44,199,614,44],"32173907","thr0wawayf00",[616,481,44,199,298,44],"32174756",[618,481,44,199,553,44],"32175037",[620,481,44,222,167,44],"32170120",[622,481,44,222,623,44],"32170357","imachine1980_",[625,481,44,222,405,44],"32171488",[627,481,48,233,628,44],"32172849","gnicholas",[630,481,44,222,631,44],"32173345","kerblang",[633,481,41,208,634,44],"32173401","nathanaldensr",[636,481,44,199,637,44],"32174953","asiachick",[639,481,44,222,640,44],"32173607","SamBam",[642,481,44,199,248,44],"32211171",[644,481,44,199,640,44],"32247719",[646,481,33,293,647,44],"32174179","hammock",[649,481,33,293,650,44],"32174209","jowday",[652,481,44,222,653,44],"32174575","djmips",[655,481,44,222,656,44],"32175361","labrador",[],[659,664,669,673,678,682],{"id":660,"name":661,"points":662,"category":663},"32096801","pledge.com",571,"dev_tools",{"id":665,"name":666,"points":667,"category":668},"32095592","Unblah",451,"productivity",{"id":670,"name":671,"points":672,"category":663},"32106126","tere",378,{"id":674,"name":675,"points":676,"category":677},"32149989","Flowful",299,"creative",{"id":679,"name":680,"points":681,"category":668},"32152600","algorithm",214,{"id":683,"name":684,"points":685,"category":686},"32121846","A compiler and VM for a simple language, in 150 lines of code",200,"languages",false,"v2",{"id":690,"name":5,"line":5,"ai":687,"cat":19,"year":12,"points":13,"comments":14,"rec":28,"s":691,"top":17,"kw":692,"maker":10,"domain":9,"tier":18,"attempts":44,"strip":693,"url":8,"mix":694,"praise":83,"objection":107,"sub":20},32167704,0.04,[23,24,25,26,27],"3105325343235130333254510343300000000513",[31,32,33,34],1793991049151]