[{"data":1,"prerenderedAt":656},["ShallowReactive",2],{"project-31903076":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":29,"sentiment":30,"stance_mix":31,"scored_comments":35,"aspects":36,"strip":47,"post_text":49,"what":50,"said":88,"maker_in_thread":204,"shape":215,"grid":220,"attempts":617,"same_week":618,"maker_shows":34,"maker_hidden":648,"reader_version":649,"card":650},"31903076","min-dalle","I stripped DALL·E Mini to its bare essentials and converted it to Torch","title","Show HN: I stripped DALL·E Mini to its bare essentials and converted it to Torch","https:\u002F\u002Fgithub.com\u002Fkuprel\u002Fmin-dalle","github.com","kuprel",1656379672,2022,663,148,27,2570,2,"legend","ai","text to image model","library",[24,25,26,27,28],"dall-e mini","text to image","pytorch","jax","image generation","curious",-0.245,{"positive":32,"neutral":33,"mixed":34,"critical":16},7,37,1,72,{"execution":37,"performance":39,"docs":41,"idea":43,"privacy":45,"design":46},{"positive":18,"critical":38},20,{"positive":34,"critical":40},0,{"positive":40,"critical":42},3,{"positive":44,"critical":18},4,{"positive":40,"critical":18},{"positive":40,"critical":40},[40,40,40,40,40,40,40,40,34,44,42,34,34,34,48,34,34,40,40,42,34,18,18,34,42,42,34,18,48,34,42,40,40,34,42,42,48,40,42,34],5,null,{"problem":51,"audience":49,"pricing":49,"motivation":49,"time_to_build":49,"stack":65,"openness":86,"stage":87,"team_size":87},{"value":52,"source":53,"evidence":54},"Running DALL·E Mini text-to-image generation with less dependency overhead, using PyTorch","inferred",[55,57,61],{"id":56},"post",{"id":58,"by":59,"text":60},"31905748","ShamelessC","They converted the original JAX weights to the format that Pytorch uses. Because JAX is still fairly new, it can be a lot easier to get Pytorch to run on e.g. CPU. I do find the…",{"id":62,"by":63,"text":64},"31906000","joshvm","To add to the sibling comment. The challenge is not converting the weights as such. Pre-trained model weights are just arrays of numbers with labels that identify which…",[66,73,78],{"value":67,"source":53,"evidence":68},"Python",[69],{"id":70,"by":71,"text":72},"31909852","dalle-world","I spun up an aws ubuntu ec2 with 2 Tesla M60. When I run python3 image_from_text.py --text='alien life' --seed=7 I get this error detokenizing image Traceback (most recent call…",{"value":74,"source":75,"evidence":76},"PyTorch","maker",[77],{"id":56},{"value":79,"source":80,"evidence":81},"JAX","commenter",[82],{"id":83,"by":84,"text":85},"31906134","phh","It defaults to JAX. To use torch, add --torch parameter","open_source","unknown",{"praise":89,"objections":105,"requests":148,"questions":173,"alternatives":189},[90,101],{"value":91,"source":80,"evidence":92},"Easier to install and run than the JAX version",[93,97],{"id":94,"by":95,"text":96},"31906023","OJFord","Look how much easier it is to install & run, people are interested in and up-voting the result, not how much work was (or wasn't) required to achieve it.",{"id":98,"by":99,"text":100},"31905660","jwitthuhn","Love this, before I only ever saw code that ran this model through jax. This seems to perform much better on my m1.",{"value":102,"source":80,"evidence":103},"Runs faster on Apple M1 than JAX code",[104],{"id":98,"by":99,"text":100},[106,118,129,140],{"value":107,"source":80,"evidence":108},"Setup errors with JAX, flax and msgpack versions",[109,110,114],{"id":70,"by":71,"text":72},{"id":111,"by":112,"text":113},"31905229","tomduncalf","Running the \"alien life\" example from the README took 30 seconds on my M1 Max. I don't think it uses the GPU at all. I couldn't get the \"mega\" option to work, I got an error…",{"id":115,"by":116,"text":117},"31905405","NhanH","Change the flax version to 0.4.2 (currently 0.5.2) will work So much for semver :(",{"value":119,"source":80,"evidence":120},"Requires a Weights & Biases account to download models",[121,125],{"id":122,"by":123,"text":124},"31905770","pja","NB: Needs a weights & biases account in order to download the models.",{"id":126,"by":127,"text":128},"31905817","mikewarot","What is wandb.ai, and Why does it keep asking for an API key? It's not listed in the requirements I've posted it as an issue",{"value":130,"source":80,"evidence":131},"Image quality is poor for some prompts",[132,136],{"id":133,"by":134,"text":135},"31906646","sAbakumoff","The results are amazingly poor. Try \"biden plays chess against napoleon\"",{"id":137,"by":138,"text":139},"31906310","JacobiX","Interesting when testing with inputs like \"Oscar Wilde photo\" or \"marilyn monroe photo\" and comparing to a Google image search. After some iterations we can have quite similar…",{"value":141,"source":80,"evidence":142},"Slow inference on CPU, about 30 seconds or more",[143,144],{"id":111,"by":112,"text":113},{"id":145,"by":146,"text":147},"31905396","klohto","The thread now has a fix. As for the GPU, it's possible to get it working with some extra steps https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fjax\u002Fissues\u002F8074 Macbook Pro M1 Pro numbers (CPU): ```…",[149,160,167],{"value":150,"source":80,"evidence":151},"Hardware and RAM requirements and inference stats",[152,156],{"id":153,"by":154,"text":155},"31904725","etaioinshrdlu","Anyone have some stats on inference time and RAM requirements? (on specific hardware)",{"id":157,"by":158,"text":159},"31903706","WheelsAtLarge","Good Job. What are the hardware requirements?",{"value":161,"source":80,"evidence":162},"Docker container support",[163],{"id":164,"by":165,"text":166},"31905196","lastdong","Now we just need to containarise it (there are a few docker python nvidia images)",{"value":168,"source":80,"evidence":169},"GPU support on Apple silicon",[170],{"id":171,"by":112,"text":172},"31905636","From reading that thread it didn't sound like GPU was fully supported yet, were you able to get it working?",[174,182],{"value":175,"source":80,"evidence":176},"What is Weights & Biases and why is an API key needed",[177,178],{"id":126,"by":127,"text":128},{"id":179,"by":180,"text":181},"31914848","mahastore","WTF is Wandb.ai? This seems like a sneaky way to get people to sign up for this wandb thingy.",{"value":183,"source":80,"evidence":184},"What is the license of generated images",[185],{"id":186,"by":187,"text":188},"31904827","JanSt","What is the license of the generated artwork?",[190,197],{"value":191,"source":80,"evidence":192},"Visions of Chaos for running text-to-image models locally on Windows",[193],{"id":194,"by":195,"text":196},"31906842","andybak","If people just want to run text to image models locally by far the easiest way I've found on Windows is to install Visions of Chaos. It was originally a fractal image generation…",{"value":198,"source":80,"evidence":199},"dalle-playground",[200],{"id":201,"by":202,"text":203},"31904887","amar-laksh","People might also like this one: https:\u002F\u002Fgithub.com\u002Fsaharmor\u002Fdalle-playground Really easy to work with",{"comments":44,"answered":40,"key":205},[206,209,212],{"id":207,"text":208},"31906404","Thanks for catching this. I just updated it so that it should work with the latest flax.",{"id":210,"text":211},"31907362","Thanks for the tip! I just updated the colab to login anonymously",{"id":213,"text":214},"31906009","Thanks for figuring this out. The problem was that the vqgan repository was being cloned to the wrong directory. The updated colab should work now",{"commenters":216,"max_depth":217,"maker_replies":44,"first_comment_mins":218,"lifespan_hours":219},104,8,110,184,[221,223,226,229,231,233,236,238,240,243,245,249,252,256,259,261,263,265,267,270,273,276,278,281,285,287,290,293,296,298,300,302,305,308,312,315,319,320,321,323,324,325,328,331,333,335,337,339,341,344,348,351,354,356,358,360,363,367,368,369,371,374,377,379,380,383,386,389,391,394,396,399,401,403,406,407,410,414,417,420,422,425,427,430,432,434,436,440,444,449,451,454,457,460,462,466,467,470,472,474,478,483,486,489,493,496,500,503,505,508,510,513,515,518,521,523,525,528,531,533,535,537,541,544,547,550,553,557,561,563,565,567,570,574,577,580,583,585,588,591,594,597,600,603,605,608,610,613],[194,40,40,222,195,40],"other\u002Fneutral\u002Fnone",[224,40,40,222,225,40],"31907516","leereeves",[227,40,40,222,228,40],"31907737","ccbccccbbcccbb",[230,40,40,222,225,40],"31907763",[232,40,40,222,228,40],"31907777",[234,40,40,222,235,40],"31909436","shreyshnaccount",[237,40,40,222,228,40],"31910349",[239,40,40,222,235,40],"31909445",[241,40,40,222,242,40],"31908114","songeater",[244,40,40,222,134,40],"31921328",[246,40,40,247,248,40],"31909477","other\u002Fpositive\u002Fnone","jquaint",[250,34,40,222,251,40],"31906026","langitbiru",[253,34,40,254,255,40],"31906731","other\u002Fcritical\u002Fnone","nudpiedo",[257,34,40,222,258,40],"31906910","slewis",[260,34,40,222,255,40],"31907577",[262,34,40,222,258,40],"31907682",[264,34,40,222,255,40],"31907764",[266,34,40,222,258,40],"31907829",[268,34,40,222,269,40],"31980526","no_identd",[271,34,40,222,272,40],"31908333","upupandup",[274,34,40,222,275,40],"31911302","harpersealtako",[277,34,40,222,272,40],"31911488",[279,34,40,222,280,40],"31911829","pmarreck",[282,34,40,283,284,40],"31913208","commenter\u002Fneutral\u002Fnone","flycaliguy",[286,34,40,283,272,40],"31920261",[288,34,40,283,289,40],"31924316","rexpop",[291,34,40,283,292,40],"31934927","cfcf14",[294,34,40,283,295,40],"31913750","khalladay",[297,34,40,222,275,40],"31991413",[70,18,34,299,71,40],"project\u002Fcritical\u002Fexecution",[301,18,34,299,280,40],"31913845",[303,42,34,299,304,40],"31904194","smcleod",[306,42,34,299,307,40],"31904244","bart3r",[309,42,42,310,311,40],"31904338","project\u002Fneutral\u002Fexecution","sam1r",[313,42,48,314,304,40],"31904405","project\u002Fpositive\u002Fexecution",[316,42,42,317,318,40],"31905058","project\u002Fneutral\u002Fperformance","julien_c",[111,42,42,317,112,40],[145,42,42,317,146,40],[322,42,42,317,112,40],"31905629",[171,42,42,317,112,40],[115,42,34,299,116,40],[326,42,40,283,327,40],"31905724","Twisol",[329,42,42,310,330,40],"31906386","lloeki",[207,42,40,332,11,34],"maker\u002Fneutral\u002Fexecution",[334,42,34,299,280,40],"31911853",[336,42,40,283,304,40],"31914167",[338,42,34,299,280,40],"31915128",[126,44,34,340,127,40],"project\u002Fcritical\u002Fdocs",[342,44,40,222,343,40],"31905855","h0mie",[345,44,42,346,347,40],"31906075","project\u002Fneutral\u002Fdocs","nl",[349,44,34,299,350,40],"31906871","Aardwolf",[352,44,40,283,353,40],"31906925","carnitine",[355,44,40,222,258,40],"31906961",[210,44,40,357,11,34],"maker\u002Fpositive\u002Fexecution",[359,44,48,357,258,40],"31907505",[361,44,34,299,362,40],"31913026","mmcnl",[364,48,42,365,366,40],"31905592","project\u002Fneutral\u002Fidea","kertoip_1",[58,48,40,283,59,40],[94,48,40,283,95,40],[370,48,40,283,59,40],"31906275",[372,48,34,299,373,40],"31906031","avhon1",[375,48,40,283,376,40],"31906058","celticninja",[83,48,40,378,84,40],"commenter\u002Fneutral\u002Fdocs",[62,48,40,283,63,40],[381,48,40,283,382,40],"31906215","mFixman",[384,48,40,283,385,40],"31906613","davidatbu",[387,48,40,283,388,40],"31906684","qayxc",[390,48,40,283,63,40],"31910407",[392,48,40,222,393,40],"31906322","nonameiguess",[395,48,40,283,388,40],"31906749",[397,48,40,222,398,40],"31907518","prionassembly",[153,400,42,317,154,40],6,[402,400,42,317,307,40],"31904979",[404,400,42,317,405,40],"31905981","chrisa",[122,32,34,340,123,40],[408,32,42,346,409,40],"31905802","lars_francke",[411,217,34,412,413,40],"31904516","project\u002Fcritical\u002Fidea","fartcannon",[415,217,42,346,416,40],"31905175","Borgz",[418,217,42,346,419,40],"31905917","capableweb",[421,217,42,346,405,40],"31905987",[423,217,34,299,424,40],"31911399","nemoniac",[426,217,42,346,280,40],"31913822",[428,217,34,340,429,40],"31906463","jjallen",[164,431,42,365,165,40],9,[157,433,42,317,158,40],10,[137,435,34,299,138,40],11,[437,435,42,438,439,40],"31907124","project\u002Fneutral\u002Fprivacy","bogwog",[441,435,34,442,443,40],"31907581","project\u002Fcritical\u002Fprivacy","l33t2328",[445,446,42,447,448,40],"31905586",12,"project\u002Fneutral\u002Fdesign","mg",[450,446,42,447,99,40],"31905641",[452,446,42,447,453,40],"31905673","01acheru",[455,446,42,447,456,40],"31905701","freemint",[458,446,42,447,459,40],"31905759","petercooper",[461,446,42,447,59,40],"31905788",[463,464,42,346,465,40],"31905462",13,"jcims",[213,464,40,332,11,34],[468,464,34,299,469,40],"31906494","DivineTraube",[471,464,42,310,465,40],"31906795",[473,464,40,332,11,34],"31907854",[475,476,34,299,477,40],"31905972",14,"bjarneh",[479,480,48,481,482,40],"31904411",15,"project\u002Fpositive\u002Fidea","godmode2019",[201,484,42,485,202,40],16,"project\u002Fneutral\u002Fnone",[487,484,34,299,488,40],"31905341","enlyth",[490,491,48,481,492,40],"31905440",17,"epicureanideal",[494,491,40,222,495,40],"31905472","cnity",[497,491,40,498,499,40],"31905553","other\u002Fneutral\u002Fidea","danielbln",[98,501,48,502,99,40],18,"project\u002Fpositive\u002Fperformance",[179,504,34,442,180,40],19,[506,38,42,317,507,40],"31907818","ramesh31",[133,509,34,299,134,40],21,[511,509,34,412,512,40],"31907678","tgv",[186,514,42,365,187,40],22,[516,514,40,222,517,40],"31905188","witheld",[519,514,40,222,520,40],"31907329","Dangeranger",[522,514,40,283,517,40],"31910573",[524,514,40,283,520,40],"31915598",[526,514,40,283,527,40],"31918350","iggldiggl",[529,514,40,222,530,40],"31924545","cuteboy19",[532,514,40,222,225,40],"31905922",[534,514,40,283,527,40],"31918262",[536,514,40,283,225,40],"31943319",[538,539,40,222,540,40],"31906609",23,"punk_ihaq",[542,539,34,299,543,40],"31906641","buf",[545,539,40,283,546,40],"31906667","mensetmanusman",[548,539,34,299,549,40],"31906642","pilotneko",[551,539,34,299,552,40],"31906660","ubertaco",[554,555,42,317,556,40],"31908825",24,"ausbah",[558,559,48,481,560,40],"31904672",25,"sydthrowaway",[562,559,40,283,154,40],"31904721",[564,559,42,365,560,40],"31905406",[566,559,42,485,59,40],"31905512",[568,559,42,485,569,40],"31905624","c0n5pir4cy",[571,572,42,485,573,40],"31904185",26,"coding123",[575,16,48,481,576,40],"31904728","daenz",[578,16,40,222,579,40],"31904770","mdp2021",[581,16,40,283,582,40],"31905104","mola",[584,16,40,222,579,40],"31905589",[586,16,40,222,587,40],"31905322","londons_explore",[589,16,40,222,590,40],"31906348","jokethrowaway",[592,16,40,222,593,40],"31904845","marban",[595,16,42,310,596,40],"31904917","_flux",[598,16,40,283,599,40],"31905496","yetihehe",[601,16,40,222,602,40],"31904919","shakna",[604,16,40,222,582,40],"31905146",[606,16,40,222,607,40],"31907223","kortex",[609,16,40,222,602,40],"31908119",[611,16,34,299,612,40],"31907391","mrtranscendence",[614,16,42,615,616,40],"31907847","project\u002Fmixed\u002Fdesign","dbddv01",[],[619,624,629,634,639,643],{"id":620,"name":621,"points":622,"category":623},"31925871","TAXSIM.app",497,"finance_crypto",{"id":625,"name":626,"points":627,"category":628},"31897430","PRQL 0.2",378,"languages",{"id":630,"name":631,"points":632,"category":633},"31910622","Legible News",320,"productivity",{"id":635,"name":636,"points":637,"category":638},"31921830","Brevity 500",193,"learning_science",{"id":640,"name":641,"points":642,"category":633},"31891608","Yboard",165,{"id":644,"name":645,"points":646,"category":647},"31905554","Teapodo",154,"creative",false,"v2",{"id":651,"name":5,"line":6,"ai":648,"cat":20,"year":13,"points":14,"comments":15,"rec":29,"s":652,"top":18,"kw":653,"maker":11,"domain":10,"tier":19,"attempts":40,"strip":654,"url":9,"mix":655,"praise":91,"objection":107,"sub":21},31903076,-0.24,[24,25,26,27,28],"0000000014311151100312213312513001335031",[32,33,34,16],1794036621894]