[{"data":1,"prerenderedAt":235},["ShallowReactive",2],{"project-37234121":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":41,"post_text":42,"what":43,"said":88,"maker_in_thread":161,"shape":169,"grid":173,"attempts":198,"same_week":199,"maker_shows":34,"maker_hidden":227,"reader_version":228,"card":229},"37234121","inference","Pip install inference, open source computer vision deployment","title","Show HN: Pip install inference, open source computer vision deployment","https:\u002F\u002Fgithub.com\u002Froboflow\u002Finference","github.com","zerojames",1692788667,2023,71,16,854,3109,28,"notable","ai","computer vision deployment","api",[21,24,25,26,27],"inference server","yolov8","docker inference","tensorrt","mixed",0.217,{"positive":31,"neutral":32,"mixed":33,"critical":34},5,6,0,3,14,{"docs":37,"idea":39},{"positive":38,"critical":38},2,{"positive":33,"critical":40},1,[31,34,33,31,34,34,31,31,40,33,40,34,34,31,40,34],"Deploying vision models is time consuming and tedious. Setting up dependencies. Fixing conflicts. Configuring TRT acceleration. Flashing (and re-flashing) NVIDIA Jetsons. A streamlined, developer-friendly solution for inference is needed.\n\nWe, the Roboflow team, have been hard at work open sourcing Inference, an open source vision deployment solution. Our solution is designed with developers in mind, offering a HTTP-based interface. Run models on your hardware without having to write architecture-specific inference code. Here's a demo showing how to go from a model to GPU inference on a video ",{"problem":44,"audience":50,"pricing":64,"motivation":65,"time_to_build":64,"stack":72,"openness":85,"stage":86,"team_size":87},{"value":45,"source":46,"evidence":47},"Deploying vision models is time consuming: dependency setup, conflicts, TensorRT configuration, and flashing Jetsons","maker",[48],{"id":49},"post",{"value":51,"source":52,"evidence":53},"Developers deploying computer vision models to production or edge devices","commenter",[54,58,61],{"id":55,"by":56,"text":57},"37235192","yeldarb","Thanks for the suggestion! Definitely agree, we’ve seen that work extremely well for Supervision[1] and Autodistill, some of our other open source projects. There’s still a lot…",{"id":59,"by":56,"text":60},"37236202","We'll work on adding this to the repo, but I'll try to summarize here. Why should I use this? If you're looking to deploy computer vision models to production (on edge devices or…",{"id":62,"by":56,"text":63},"37236813","Inference operates at a higher layer of abstraction (tasks & architectures vs weights), targets a different audience (software engineers vs deep machine learning experts), and is…",null,{"value":66,"source":46,"evidence":67},"Reduce friction of deploying vision models; internal Roboflow project open sourced",[68,69],{"id":49},{"id":70,"by":56,"text":71},"37241209","Yes, but also not easily right now. We’ll be adding more supported ways to load models soon, but right now things largely assume using our model registry (and API Key is how it…",[73,77,81],{"value":74,"source":46,"evidence":75},"Docker",[76],{"id":49},{"value":78,"source":46,"evidence":79},"TensorRT",[80],{"id":49},{"value":82,"source":52,"evidence":83},"Python",[84],{"id":70,"by":56,"text":71},"open_source","launched","company",{"praise":89,"objections":112,"requests":131,"questions":143,"alternatives":151},[90,105],{"value":91,"source":52,"evidence":92},"Useful launch and tool, recommended by users",[93,97,101],{"id":94,"by":95,"text":96},"37237281","tlack","Love & recommend Roboflow. Thanks for releasing this yall!",{"id":98,"by":99,"text":100},"37236201","pkiv","Looks great! Congrats on the launch.",{"id":102,"by":103,"text":104},"37235435","jvanillaaaa","This is sick",{"value":106,"source":52,"evidence":107},"Video in README was impressive",[108],{"id":109,"by":110,"text":111},"37244409","noman-land","Thanks. The video blew my mind.",[113,123],{"value":114,"source":52,"evidence":115},"Unclear what the project is and why to use it, docs lacked WHY",[116,120],{"id":117,"by":118,"text":119},"37235427","Aspos","Interested, but after spending 2 min reading it I am still not sure what I am looking at exactly.",{"id":121,"by":118,"text":122},"37235584","Genuine feedback: I guess your readme.md describes the HOW well, but does not touch the WHY. Why should I use this? Why not?",{"value":124,"source":52,"evidence":125},"Requires Roboflow API key to run",[126,130],{"id":127,"by":128,"text":129},"37240967","thangngoc89","Is it possible to run this without supplying Roboflow API key? Why the requirement?",{"id":70,"by":56,"text":71},[132,139],{"value":133,"source":52,"evidence":134},"Add images or video to README",[135],{"id":136,"by":137,"text":138},"37235095","nl","Looks interesting. Missing a trick by not having some images in the README though. Especially since this post here is where I found out it can run CLIP instead of the readme.",{"value":140,"source":52,"evidence":141},"Support running without Roboflow API key",[142],{"id":127,"by":128,"text":129},[144],{"value":145,"source":52,"evidence":146},"How does it compare to NVIDIA Triton?",[147],{"id":148,"by":149,"text":150},"37236459","claytonjy","How does this compare to Nvidia Triton? It's also a dockerized inference server capable of running ONNX, TRT, and many other compiled models. Automatic request queuing, HTTP &…",[152,157],{"value":153,"source":52,"evidence":154},"NVIDIA Triton",[155,156],{"id":148,"by":149,"text":150},{"id":62,"by":56,"text":63},{"value":158,"source":52,"evidence":159},"DeepStream",[160],{"id":59,"by":56,"text":60},{"comments":38,"answered":33,"key":162},[163,166],{"id":164,"text":165},"37237171","I have added in a video showing inference into the README -- thank you for the feedback!",{"id":167,"text":168},"37235535","Thank you for your comment! Inference is all about reducing the friction associated with deployment. Set up a Docker container (which works on a range of architectures, with TRT…",{"commenters":170,"max_depth":171,"maker_replies":38,"first_comment_mins":172,"lifespan_hours":15},10,4,111,[174,176,178,179,180,182,184,186,187,189,190,191,192,194,195,197],[136,33,31,175,137,33],"project\u002Fpositive\u002Fdocs",[55,33,34,177,56,33],"maker\u002Fneutral\u002Fdocs",[164,33,33,177,11,40],[109,33,31,175,110,33],[148,40,34,181,149,33],"project\u002Fneutral\u002Fidea",[62,40,34,183,56,33],"maker\u002Fneutral\u002Fidea",[94,38,31,185,95,33],"project\u002Fpositive\u002Fnone",[98,34,31,185,99,33],[117,171,40,188,118,33],"project\u002Fcritical\u002Fdocs",[167,171,33,177,11,40],[121,171,40,188,118,33],[59,171,34,183,56,33],[193,171,34,181,56,33],"37235979",[102,31,31,185,103,33],[127,32,40,196,128,33],"project\u002Fcritical\u002Fidea",[70,32,34,183,56,33],[],[200,205,209,214,218,222],{"id":201,"name":202,"points":203,"category":204},"37190743","flipreps",373,"creative",{"id":206,"name":207,"points":208,"category":204},"37194128","Just intonation keyboard",331,{"id":210,"name":211,"points":212,"category":213},"37188659","Rivet",327,"infrastructure",{"id":215,"name":216,"points":217,"category":204},"37188102","Avatartion",162,{"id":219,"name":220,"points":221,"category":213},"37224148","FlakeHub",154,{"id":223,"name":224,"points":225,"category":226},"37234887","Fast vector similarity",141,"data",false,"v2",{"id":230,"name":5,"line":6,"ai":227,"cat":20,"year":13,"points":14,"comments":15,"rec":28,"s":231,"top":18,"kw":232,"maker":11,"domain":10,"tier":19,"attempts":33,"strip":233,"url":9,"mix":234,"praise":91,"objection":114,"sub":21},37234121,0.22,[21,24,25,26,27],"5305335510133513",[31,32,33,34],1794036640562]