[{"data":1,"prerenderedAt":330},["ShallowReactive",2],{"project-21558343":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":43,"post_text":45,"what":46,"said":72,"maker_in_thread":147,"shape":149,"grid":154,"attempts":290,"same_week":291,"maker_shows":41,"maker_hidden":322,"reader_version":323,"card":324},"21558343","2019","Autopilot written in Keras for Self Driving Cars","title","Show HN: Autopilot written in Keras for Self Driving Cars","https:\u002F\u002Flittlemountainman.github.io\u002F2019\u002F11\u002F27\u002Fselfdrivingfun\u002F","littlemountainman.github.io","autonomydriver",1574004942,2019,144,58,304,2195,14,"notable","ai","self-driving model","other",[24,25,26,27,28],"keras","self driving","autopilot","neural network","driving model","curious",0.048,{"positive":32,"neutral":33,"mixed":34,"critical":35},9,16,1,5,31,{"execution":38,"design":40,"docs":42},{"positive":39,"critical":34},0,{"positive":39,"critical":41},2,{"positive":39,"critical":39},[44,35,35,39,44,39,39,39,44,39,34,39,39,44,39,44,39,39,39,39,35,35,44,39,44,35,35,34,44,35,34,39,34,39,34,44,44,44,35,35],3,null,{"problem":47,"audience":45,"pricing":45,"motivation":45,"time_to_build":45,"stack":59,"openness":69,"stage":70,"team_size":71},{"value":48,"source":49,"evidence":50},"Predicting driving behaviour from video to steer and control a car","inferred",[51,55],{"id":52,"by":53,"text":54},"21564027","pathsjs","> At the current state of my model the model basically just clones the human driver as good as possible. That means the the amount of brake is higher in curves I read in another…",{"id":56,"by":57,"text":58},"21562178","littlemtman","No, I am not testing this on real roads only testing with videos so far. So no real cars only videos in cars",[60],{"value":61,"source":62,"evidence":63},"Keras","commenter",[64,68],{"id":65,"by":66,"text":67},"21561670","maldeh","Whoa, before we start pulling out the pitchforks, I think this may be terribly distorting OP's words and intent. - My reading of it was that OP was in awe at the immense scale of…",{"id":52,"by":53,"text":54},"unknown","prototype","solo",{"praise":73,"objections":86,"requests":105,"questions":117,"alternatives":135},[74],{"value":75,"source":62,"evidence":76},"Impressive work for a young developer",[77,78,82],{"id":52,"by":53,"text":54},{"id":79,"by":80,"text":81},"21561588","nikodunk","Great job! This is amazing. Love how I’m you used off-the-shelf components. Congrats and keep going!!",{"id":83,"by":84,"text":85},"21565466","rpmisms","Holy crap. First: you're awesome. Keep it up. Second: anyone else think this is a great example of how autonomous driving is a last-mile problem?",[87,97],{"value":88,"source":62,"evidence":89},"Neural network alone is unsafe for safety-critical driving",[90,94],{"id":91,"by":92,"text":93},"21562919","devit","Seems really scary to exclusively use a neural network for safety-critical tasks like this, without having an explicit method guaranteeing safety. You can't prove that a trained…",{"id":95,"by":57,"text":96},"21563900","The risk is extremely high.",{"value":98,"source":62,"evidence":99},"Braking before curves is not how most drivers behave",[100,101],{"id":52,"by":53,"text":54},{"id":102,"by":103,"text":104},"21564898","hugey010","You are conflating effective driving habits with actual driving habits. I'd say less than 5% of drivers in the states properly brake before curves and accelerate through. Hell,…",[106,113],{"value":107,"source":62,"evidence":108},"Add temporal context from previous frames",[109],{"id":110,"by":111,"text":112},"21561824","czr","good start op. a good next step is add in temporal context (previous frame information) so model may resolve ambiguous cases. see e.g. breakdown of comma.ai openpilot model here…",{"value":114,"source":62,"evidence":115},"Lead with demos and failure cases",[116],{"id":110,"by":111,"text":112},[118,125],{"value":119,"source":62,"evidence":120},"How is overfitting handled given held-out test data",[121],{"id":122,"by":123,"text":124},"21563447","billsmithaustin","How big of a risk is overfitting for these kinds of models? I think common practice is to hold back some of the data for testing.",{"value":126,"source":62,"evidence":127},"Why ELU activation works better than ReLU",[128,132],{"id":129,"by":130,"text":131},"21561751","syntaxing","You do something similar in Udacity's SDC nanodegree but you use their simulator rather than a new car. Interestingly through trial and error on my project, ELU activation…",{"id":133,"by":130,"text":134},"21562405","Thanks for the answer! I also used Relu with no luck either. I figured Relu would work better than ELU but it didn't work for this application. Would you happen to know why ELU…",[136,140],{"value":137,"source":62,"evidence":138},"comma.ai openpilot",[139],{"id":110,"by":111,"text":112},{"value":141,"source":62,"evidence":142},"Tesla Autopilot",[143],{"id":144,"by":145,"text":146},"21561478","daenz",">A few days ago @karpathy presented their workflow with PyTorch and also said some numbers, to train the Autopilot system with all it neural networks you would have to spend…",{"comments":39,"answered":39,"key":148},[],{"commenters":150,"max_depth":151,"maker_replies":39,"first_comment_mins":152,"lifespan_hours":153},27,6,659,47,[155,157,160,162,166,168,171,173,175,177,180,183,186,188,190,192,195,198,201,204,207,210,212,215,217,220,222,225,227,229,230,232,235,237,238,240,242,243,245,246,248,252,254,256,259,262,265,268,270,271,273,276,278,281,283,284,286,288],[52,39,44,156,53,39],"project\u002Fmixed\u002Fexecution",[158,39,35,159,57,39],"21564498","maker\u002Fpositive\u002Fnone",[161,39,35,159,84,39],"21565474",[163,39,39,164,165,39],"21564877","other\u002Fneutral\u002Fnone","mannykannot",[102,39,39,167,103,39],"commenter\u002Fneutral\u002Fnone",[169,39,39,167,170,39],"21565341","AllanHoustonSt",[110,34,44,172,111,39],"maker\u002Fneutral\u002Fnone",[174,34,39,167,130,39],"21561949",[176,34,39,167,57,39],"21563049",[178,34,39,179,111,39],"21564521","commenter\u002Fneutral\u002Fdocs",[181,34,39,164,182,39],"21563051","jacquesm",[184,34,44,185,57,39],"21563939","project\u002Fneutral\u002Fexecution",[187,34,39,167,182,39],"21564023",[189,34,39,167,57,39],"21564572",[91,41,34,191,92,39],"project\u002Fcritical\u002Fdesign",[193,41,39,164,194,39],"21563010","_Microft",[196,41,39,164,197,39],"21563055","solarkraft",[199,41,39,164,200,39],"21564619","gnode",[202,41,39,164,203,39],"21563583","chronolitus",[205,41,44,206,57,39],"21563040","project\u002Fneutral\u002Fdesign",[208,41,39,167,209,39],"21563698","yitchelle",[211,41,44,206,57,39],"21563906",[213,41,39,167,214,39],"21564056","lnsru",[216,41,39,167,57,39],"21564525",[218,41,39,167,219,39],"21564494","TulliusCicero",[221,41,39,167,57,39],"21564531",[223,41,39,167,224,39],"21565514","mmcwilliams",[226,41,39,167,209,39],"21573278",[228,41,39,167,200,39],"21564681",[79,44,35,159,80,39],[231,44,35,159,57,39],"21562229",[129,233,44,234,130,39],4,"maker\u002Fneutral\u002Fexecution",[236,233,44,185,57,39],"21562228",[133,233,39,167,130,39],[239,233,39,167,57,39],"21563914",[241,233,44,234,111,39],"21564542",[83,35,35,159,84,39],[244,35,35,159,57,39],"21565616",[122,151,44,185,123,39],[95,151,34,247,57,39],"project\u002Fcritical\u002Fexecution",[249,250,44,172,251,39],"21565062",7,"verumn",[253,250,44,172,57,39],"21565279",[255,250,35,159,251,39],"21565336",[144,257,34,258,145,39],8,"project\u002Fcritical\u002Fnone",[260,257,39,167,261,39],"21561515","serf",[263,257,39,164,264,39],"21561549","FpUser",[266,257,34,191,267,39],"21561608","WrtCdEvrydy",[269,257,39,167,57,39],"21562211",[65,257,39,167,66,39],[272,257,34,258,145,39],"21562071",[274,257,44,275,57,39],"21562198","maker\u002Fneutral\u002Fdocs",[277,257,44,275,182,39],"21562777",[279,257,44,275,280,39],"21562885","mft_",[282,257,44,172,66,39],"21562207",[56,257,44,172,57,39],[285,32,35,159,182,39],"21562774",[287,32,44,172,57,39],"21563043",[289,32,35,159,182,39],"21563066",[],[292,297,302,307,312,317],{"id":293,"name":294,"points":295,"category":296},"21559139","Chatwoot",417,"web_saas",{"id":298,"name":299,"points":300,"category":301},"21561613","BrowserGap",253,"security_privacy",{"id":303,"name":304,"points":305,"category":306},"21587935","EasyDB",158,"data",{"id":308,"name":309,"points":310,"category":311},"21545291","Scholarref tools",127,"productivity",{"id":313,"name":314,"points":315,"category":316},"21547369","Faster UTF-8 validator",122,"languages",{"id":318,"name":319,"points":320,"category":321},"21533921","Bamboolib",119,"dev_tools",false,"v2",{"id":325,"name":5,"line":6,"ai":322,"cat":20,"year":13,"points":14,"comments":15,"rec":29,"s":326,"top":18,"kw":327,"maker":11,"domain":10,"tier":19,"attempts":39,"strip":328,"url":9,"mix":329,"praise":75,"objection":88,"sub":21},21558343,0.05,[24,25,26,27,28],"3550300030100303000055303551351010133355",[32,33,34,35],1794036583048]