[{"data":1,"prerenderedAt":315},["ShallowReactive",2],{"project-47995037":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":36,"aspects":37,"strip":44,"post_text":46,"what":47,"said":68,"maker_in_thread":164,"shape":176,"grid":180,"attempts":277,"same_week":278,"maker_shows":39,"maker_hidden":307,"reader_version":308,"card":309},"47995037","ml-sharp-web","Apple's SHARP running in the browser via ONNX runtime web","title","Show HN: Apple's SHARP running in the browser via ONNX runtime web","https:\u002F\u002Fgithub.com\u002Fbring-shrubbery\u002Fml-sharp-web","github.com","bring-shrubbery",1777799696,2026,185,46,328,4366,8,"notable","ai","image to 3d gaussian splat","web_app",[24,25,26,27,28],"onnx runtime web","gaussian splatting","image to 3d","browser ml","sharp","mixed",-0.011,{"positive":32,"neutral":33,"mixed":34,"critical":35},3,7,1,4,15,{"idea":38,"execution":40,"performance":42,"docs":43},{"positive":39,"critical":34},2,{"positive":34,"critical":41},0,{"positive":41,"critical":32},{"positive":41,"critical":41},[41,41,41,41,41,41,45,41,41,41,41,32,41,41,45,32,41,34,41,41,41,41,41,41,41,41,32,41,34,41,41,45,32,34,41,32,34,32,41,32],5,null,{"problem":48,"audience":46,"pricing":46,"motivation":46,"time_to_build":46,"stack":59,"openness":66,"stage":67,"team_size":66},{"value":49,"source":50,"evidence":51},"Turns a single photo into a 3D Gaussian splat scene","commenter",[52,56],{"id":53,"by":54,"text":55},"47996508","amelius","I don't like that it uses only a single photo. This means it is going to make up a lot of stuff. E.g. if I show it a photo of a poster, then it will make that poster 3D. With…",{"id":57,"by":11,"text":58},"47999027","Yeah I completely agree, but I think this model solves a different problem. AFAIK it's specifically there for the case where you only have one photo, but still need a 3D gaussian…",[60],{"value":61,"source":62,"evidence":63},"ONNX Runtime Web","maker",[64],{"id":65},"post","unknown","launched",{"praise":69,"objections":88,"requests":122,"questions":137,"alternatives":153},[70,81],{"value":71,"source":50,"evidence":72},"Impressive browser tech showcase and fast inference",[73,77],{"id":74,"by":75,"text":76},"47999648","echelon","> inference itself is a few seconds on a recent Mac This is impressive as hell Very cool demo. It works in about ~9 seconds on my machine. A few asks if you're going to devote…",{"id":78,"by":79,"text":80},"47996007","javier2","Did not work in Firefox on Linux, but it runs on Chrome. Have to admit, I dont get it. I tried it with 3 landscape photos I have and the results were nowhere close to the results…",{"value":82,"source":50,"evidence":83},"Mesmerising spatial viewing of photos in VR",[84],{"id":85,"by":86,"text":87},"47996413","andybak","I vibecoded a simple web app using Sharp that allowed be to quickly browse any local image folder and view them as \"almost\" volumetric 3d scenes in a VR headset. I precomputed…",[89,107,114,118],{"value":90,"source":50,"evidence":91},"Model is about 2.4 to 2.5GB and crashes browsers or runs out of memory",[92,96,100,103],{"id":93,"by":94,"text":95},"47996708","herpdyderp","Loading the model crashes my browser tab from memory usage :\u002F",{"id":97,"by":98,"text":99},"47997752","jeroenhd","What are the requirements for running this? Chrome throws a whole bunch of \"out of memory\" errors into the console when I try to execute these. I'm guessing 4GiB of VRAM is not…",{"id":101,"by":98,"text":102},"48006198","My laptop's GPU has 4GiB of VRAM but it still failed to allocate enough memory. Seems like I'll have to pass on demos like these until someone figures out a way to use less…",{"id":104,"by":105,"text":106},"47999843","anentropic","It froze up my computer, had to hard-boot lol (16MB M1 Macbook, Chrome)",{"value":108,"source":50,"evidence":109},"Single photo makes up depth that is not real",[110,111],{"id":53,"by":54,"text":55},{"id":112,"by":54,"text":113},"47996798","Maybe, but what is wrong with wanting real depth instead of \"made up depth\"? One extra photo mostly solves that.",{"value":115,"source":50,"evidence":116},"Results differ from demo; poor on some photos",[117],{"id":78,"by":79,"text":80},{"value":119,"source":50,"evidence":120},"Did not work in Firefox on Linux",[121],{"id":78,"by":79,"text":80},[123,133],{"value":124,"source":50,"evidence":125},"Lower memory or compressed model for low-RAM machines",[126,129],{"id":127,"by":11,"text":128},"47999247","Yeah, I think you need at least 8gb ram unfortunately, but I tested it only on a 32gb M2, so 8gb might also not be enough. I might create a compressed version of the model, that…",{"id":130,"by":131,"text":132},"48017154","mattbaconz","2.4GB in browser is crazy. I’m curious whether quantization works here or if the model quality falls apart too much.",{"value":134,"source":50,"evidence":135},"Full orbital camera and double-click drag to move camera",[136],{"id":74,"by":75,"text":76},[138,142,149],{"value":139,"source":50,"evidence":140},"What are the memory requirements?",[141],{"id":97,"by":98,"text":99},{"value":143,"source":50,"evidence":144},"Does it use WASM and WebGPU?",[145],{"id":146,"by":147,"text":148},"48002824","sroussey","Does it use both WASM and WebGPU?",{"value":150,"source":50,"evidence":151},"Does quantization work without losing quality?",[152],{"id":130,"by":131,"text":132},[154,160],{"value":155,"source":50,"evidence":156},"Original 3D Gaussian Splatting, Polycam, Luma, Postshot for multi-photo",[157],{"id":158,"by":11,"text":159},"47999288","For multi-photo, the go-to is still the original 3D Gaussian Splatting (Kerbl et al., 2023) — most consumer tools like Polycam, Luma, and Postshot wrap that under the hood.",{"value":161,"source":50,"evidence":162},"HY-World 2.0 and WorldLabs Marble",[163],{"id":74,"by":75,"text":76},{"comments":165,"answered":35,"key":166},10,[167,170,173],{"id":168,"text":169},"47998991","Interesting, what use cases are you using onnx for btw?",{"id":171,"text":172},"47998976","Yeah it's pretty cool what a 2gb NN can do from a single image",{"id":174,"text":175},"47999000","Nice, would love to see it, feel free to link it here \u003C3",{"commenters":177,"max_depth":45,"maker_replies":165,"first_comment_mins":178,"lifespan_hours":179},21,155,54,[181,185,188,190,192,195,197,198,202,205,207,208,211,214,217,219,221,223,225,228,229,231,234,236,238,240,244,245,247,249,252,253,257,258,260,261,263,264,269,272,275],[182,41,41,183,184,41],"47997730","other\u002Fneutral\u002Fnone","exabrial",[186,41,41,183,187,41],"47998346","vunderba",[168,41,41,189,11,34],"maker\u002Fneutral\u002Fnone",[191,41,41,183,187,41],"47999253",[193,41,41,183,194,41],"48003927","Tsarp",[171,41,41,196,11,34],"project\u002Fpositive\u002Fidea",[85,34,45,196,86,41],[199,34,41,200,201,41],"47997284","commenter\u002Fpositive\u002Fnone","SpyCoder77",[203,34,41,204,86,41],"47997336","commenter\u002Fneutral\u002Fnone",[206,34,41,204,86,41],"48023360",[174,34,41,189,11,34],[209,34,32,210,86,41],"48023351","project\u002Fneutral\u002Fnone",[212,39,41,183,213,41],"47996194","kodablah",[215,39,41,216,11,34],"47999042","project\u002Fneutral\u002Fperformance",[74,32,45,218,75,41],"project\u002Fpositive\u002Fexecution",[78,35,32,220,79,41],"project\u002Fmixed\u002Fexecution",[222,35,41,216,11,34],"47999283",[53,45,34,224,54,41],"project\u002Fcritical\u002Fidea",[226,45,41,227,86,41],"47996657","commenter\u002Fneutral\u002Fidea",[112,45,41,227,54,41],[230,45,41,227,86,41],"47997324",[232,45,41,227,233,41],"47996968","deanva",[235,45,41,204,86,41],"47997315",[237,45,41,204,233,41],"48001770",[57,45,41,239,11,34],"project\u002Fneutral\u002Fidea",[241,242,41,183,243,41],"47999850",6,"parentheses",[97,33,32,216,98,41],[246,33,41,216,11,34],"47999073",[101,33,34,248,98,41],"project\u002Fcritical\u002Fperformance",[250,18,41,183,251,41],"47996977","vessenes",[158,18,41,183,11,34],[254,255,45,196,256,41],"48003143",9,"david_mchale",[130,165,32,216,131,41],[93,259,34,248,94,41],11,[127,259,41,216,11,34],[262,259,32,216,213,41],"47999560",[104,259,34,248,105,41],[265,266,32,267,268,41],"47997406",12,"project\u002Fneutral\u002Fdocs","andruby",[270,266,41,271,11,34],"47999225","project\u002Fneutral\u002Fprivacy",[273,266,32,267,274,41],"48005444","utopiah",[146,276,32,216,147,41],13,[],[279,284,289,293,298,303],{"id":280,"name":281,"points":282,"category":283},"47972511","WhatCable",566,"hardware",{"id":285,"name":286,"points":287,"category":288},"48038257","Hallucinopedia",317,"creative",{"id":290,"name":291,"points":292,"category":288},"48026342","PaletteInspiration",215,{"id":294,"name":295,"points":296,"category":297},"48037724","Tilde.run",205,"dev_tools",{"id":299,"name":300,"points":301,"category":302},"48038019","Templatical",169,"productivity",{"id":304,"name":305,"points":306,"category":20},"47990708","hnup.date",168,false,"v2",{"id":310,"name":5,"line":6,"ai":307,"cat":20,"year":13,"points":14,"comments":15,"rec":29,"s":311,"top":18,"kw":312,"maker":11,"domain":10,"tier":19,"attempts":41,"strip":313,"url":9,"mix":314,"praise":71,"objection":90,"sub":21},47995037,-0.01,[24,25,26,27,28],"0000005000030053010000000030100531031303",[32,33,34,35],1794036700300]