How to Deploy your Computer Vision Models on Raspberry Pi using Roboflow Inference

hey in this video you will see we will guide you through the processing of the exporting your Trend model from Robo flow setting up youri transferring the model and running in front where you have a beginner or an experence developer this tutorial has something for everyone interest in Ai and iot do not miss out the hand on guide to bring your AI project to life on Rasberry pie so guys let's start from the beginning let's start from a sketch in this video you will see we are going to exploring how to uh real time deploy uh your own customized model and resper pie so let's start in front of you uh I am recording two types of uh one is my laptop screen second one is my uh Reser screen so Reser screen will be capture my phone and uh my screen will be captured on my on laptop so you can see uh I am uh log in with uh Team your access and you can see we are going to exploring uh real time uh deduction of uh uh my AI model so here you can see uh we are capturing one one uh one by one uh in front of you uh model of the inference library and uh robot flow live uh robot flow model and you can see we are uh one is the Caesar second one is paper and third one is Rogue and we will deploy uh simple model of the inference one example you will create your own way of the example will be deployed on on rasper P and webcam you will uh create your own camera just like put your customized model and you will detect anything okay so in my previous blog and videos uh I am Deploy on the thief deduction and now this logic will be changed uh we are capturing the uh Hardware plus software both sides in front of you and uh it's accuracy and you also see the and uh deduction part will be captured in front of you it's real time so uh uh first of all you subscribers and uh here our website you easily get the notifications and now uh I am closed the all process and we are going to step by step first you need to con understand the concept second one we will uh going to exploring one by one point and uh do not worry we this video will be uh Med I am step by step okay how to get access in in front of you I am getting access and with the team here and how to install all dependency in Ras pie let me show you and first of all you will see uh we are going to roof flow uh customized model and one example I am already uh provide the my gith account so you will get the uh this all whole story will be linked in my description uh so I am going to step by step do not worry first you need to uh code and uh inference Library installation part so I am going to first of all you can see inference uh Robo flow okay uh object dection on uh using inference Library okay so here uh I am going to my previous report of the inference library and uh you will see I am just in copy this code and paste my uh on the resp by side first we need to understand uh how to installation part will be covered and uh I am going to Robo flow here the pro universe and also Robo flow project okay so and documentation okay so uh we need to deploy uh inference Library here inference API okay so do not worry I am going to step by step and I'm copy from here okay so document Robo flow okay so here API inference you also get from the step by-step Library type uh documentation in from the robo flow and uh I am showing to you inference uh installation okay so here you see the documentation and uh uh Quicky start you will see the Quicky start uh now we will implement the here inference okay so you can see uh here the some points of the inference what is the inference and how to deply already I am uh showing you installation part will be covered and this video and already creating much videos in with the inferences concept how to deploy your own inference model and now this time we will Deploy on the respberry pie okay so I am open my terminal okay so you can see we are going to step by step uh you need to have ubu ubu resper P I will show you which type of the uban 2 you need but this time we will going to cover uh installation part of the inference and uh here you can see just one command pip install inference all dependency will be uh automatically uh installation part is done so you need to also resy for yubu here you download from here okay after you download you can see the instruction of the res Peri how to uh in your boot uh res perie so it's the same and you download it you if after download the this version I'm already download and installation part is done imaginer of the r p you will boot this uh file on the RAS P from here okay you need to uh just download I make user I am already download and uh I will show you uh imager here I am open it okay I'm open the imaginer and uh this time I'll show this this point will be is very uh you mention your card and res Peri 4 model I am using the res 4 and I'm using other General operating system okay ubu you will uh boot the from download the ubun 2 and uh here back and customize use customize okay so you will easily to access your uh already download the UB 2 model and Bo put the model from here and also you it's very simple go to the other journal and you want to choose this one okay so here there is by 3 4 S and zero you use this one sorry server not server desktop version let me correct and this desktop version okay this part will be covered then choose the storage your mention your car card my C is already in my R pie so this two points will be covered and uh now we will going to the resper pie side and uh I am choosing one example uh of the this time inference of the with be hand will be detect uh three four types of categories will be uh D so here I will show you inference uh model which I get from the here Universe okay so after you can see we need to one API key also so I am going to uh here setting and API key here okay so do not show your code just copy and uh in my case do not worry I am already content created so not issue my side I deleted and uh you will going to stay by step in this video so just copy and paste in your rasperry pie so now in this time we need to understand uh how to copy and paste all things and uh uh when you install the this library is going to taking uh around uh too much time also depend on the internet okay so you can see just uh in front of you I we cannot copy the right path just copy okay so here paste okay so again creating issue I do not know why creating issue so I am already put the this key in my rber p uh in the B file so when I am when I open the terminal it's automatically recognize my key so do not worry we will uh going to stay by step and I am also put the keyboard so uh sud sudo sud Sudo is the here recognize the ubu Nano means uh you will will create your uh in the Nano part uh script side here the best script means when you open your uh first first need to understand B when you open your res Peri which type of categories will be open automatically you will mention here and you can see we are same command same URL put in the b files okay so I just copy from from here okay and in front of you I will past here okay it's now my terminal is ready to use the inference okay now need to understand how to I am put the inference uh this type of a script means inference library and model so you can see uh I am get the one algorithm from the inference Library so you you this report will be in front of you you will get the easily from my GitHub report okay so Link in description now we need to uh create the this demo okay and uh you can also see the how to get the new model so I am totally uh mentioned the robo flow so because I am uh deploy customized model from the inference Library which Robo flow already provide the open source uh so in front of you you can see here demo uh sorry test go to the test here you can see here we are going to the in front of you uh here webcam example okay so uh just copy this code and uh this code will be provided in my g account so do not worry so copy this code and past in your rest verify script so in front of you I mention the camera and model you will replace your model from the uh I will show you you will see we if you want to this type of model you will just go to the API and uh replace the uh first you need to train your own model so you do not need to this type of but this concept you will need to understand you will just copy this link and paste in your algorithm side here in this algorithm side okay so this time my side is okay so need to understand this libraries uh get from the inference and initialize uh a pipeline of the object uh model ID and uh video uncore reference you will mention your video and then you will see the your function okay so now need to close it uh we do not need to save it again or now need to open uh terminal I'm just right click and uh left click so in front of you uh I'm going to the desktop directory and need to run this code okay so first you need to understand we are going to Python 3 version not the uh other version Python 3 version Python 3 test.py so I hope so it's now will be uh accurate data show mean means the deduction show um I will deploy roof flow customized model in my raspber pi and live camera so we will create as soon as uh our own camera and you will also see the we are going to the deploy simple example if you have rasper P you content creator and you also learner you will try it so we are going to step by step okay so in front of you I am captured the my camera uh and uh uh Rock okay sorry paper so you will see the three type of categories and uh accuracy detection part frame per seconds and I will not mention here the frame per second but I make sure that around 60 frame per seconds will be showing you so detection part is done and accurate customized model in front of you will Beed and uh here uh I am mentioned the all signs then deduction part in front of you will be showing you if you want to create your own model and you will detect uh you can do it and uh so in this video you you will learn how to train your own sorry customized model in your Rasberry P so hope you like it keep supporters and uh this video is informational so share with your friends and uh hope you like it and keep support us see you in next video

Channel: Pyresearch

Views: 959

Keywords: Education, Pyresearch, Pyresearch - YouTube, how to Grow, money, facebook app, facebook, how to set goal, computervision, machinelearning, opencv, deeplearning, artificialintelligence, datascience, nlp, chatgpt4, ComputerVision, Roboflow, RaspberryPi, Inference, MachineLearning, DeepLearning, TechTutorial, ModelDeployment, ArtificialIntelligence, ComputerScience, EmbeddedSystems, ImageRecognition, ObjectDetection, EdgeComputing, DeveloperCommunity, TrendingNow, ViralVideo, InternetSensation, WatchThis, GoingViral, IOT, iot

Id: m7J3iNuEFs4

Length: 16min 57sec (1017 seconds)

Published: Sun Mar 31 2024