4. Downloads Data & Configure Path Yolo v7 | Object Detection | Computer Vision

hello everyone welcome to this video in this video I'll be showing you like how to download your custom data sets from this URL I had uploaded my data set in my GitHub repository now I will be showing you like how to download this data set in this collab notebook okay and how to unzip that and how to prepare this data for our model training okay and I had already shown you like how to install this uh like yellow V7 okay inside this collab notebook okay so this is the code actually I will share this notebook with you okay so that you don't have to create this notebook so let's run one by one okay first of all I'm doing the installation and all okay so this is the same command I am writing but I have kept inside uh uh one cell okay all the comments I have kept inside of oneself and if you see it's installing torch Library okay so it is utilizing torch 120 the 10.0 okay now let's wait so guys our installation is done okay so now this is the code for this uh downloading the data set from the URL okay so the same URL I had copied from here like click on this download and copy the link at address and paste this link here okay so let's just link here the same link I've just pasted okay so once you have done so this is the code uh for this so this is the code for this so what I'm doing I'm downloading this data uh like uh I'm saving it a data underscore yellow v7.3 and then I'm doing the unzipping operation okay of this data and I'm deleting this ZIP file okay so once it is done and here I am using call command Okay Carl will actually download the data if you know like if you don't know Carl is a Linux command okay you can use call command to download anything okay so now let's run run this file so now let's run this cell see it has downloaded my data so see all the image file and all the level five I have okay inside this collab okay right now so now what I need to do uh so I think you remember we have kept to file empty like one is friend.txt if you see here and one is like well.txt okay so now we need to uh fill this file okay with some uh like logic so what logic actually we need to do so here actually you need to prepare your image path okay in this txt file because yellow V7 takes the emails uh like path okay so if you see this is my images okay so inside Trend I have all my images so if I open this one uh so this is my image I suppose okay and if I just copy path here copy path and if I suppose paste it here so you can see this is the path of that images Okay so this path actually I need to mention inside inside the txt file I have created not only one path I have to save all the path okay so this is the technique okay in this yellow V7 you can talk about okay so now let's run one by one so for that first of all what I'm doing I'm writing a logic here so I am first of all taking my training images folder okay this is my training images folder you can see uh this is the training uh training images folder location okay I'm just copying the path and I'm testing here okay then I'm taking my validation folder path okay this is the validation for that path okay now let's run it now I'm entering inside my content okay so this is the content and this is the logic actually for training images first of all what I'm doing I'm just opening this train.txt file in appending mode then I'm doing this uh OS dot list ad okay on top of this folder so it will return all the list of the images okay from this folder okay all the list of the images and I'm going through a I'm applying your for Loop okay I'm taking all all the images one by one and I'm joining the path okay and I'm saving the path information now if I run it now if I open my trend.txt now you will observe all the path I have saved here okay so this is the logic of this code okay so it will actually take all the image part and it will save inside this trend.txt folder and similar wise you you need to also do for the validation set okay so this is the code for that so I am opening this welder testing appending mode and I'm doing the same okay now if I open my validation.txt so see all the validation image has been saved here okay so now once it is done you need to uh like do it you need to make a data data configuration file okay so so for that you need to enter inside yellow V7 there is a folder called Data okay so this is the data so what I need to do you need to make a copy of this coco.aml okay coco.aml you need to make a copy okay so this is the command for this so it will just take this coco.aml until make a copy in the same folder but with the name of this custom.emf okay so now let's run it okay now if I refresh here see custom.yaml has been created okay now if I open this custom.yaml okay so now here actually you need to make some modification okay so if you see it has been trained with the Coco data set that's why the number of classes was 80 okay so these are the classes they have Trend and this is the like train location validation location test location okay so here we are not going to use that we'll be using our own uh like Uh custom uh here we'll be using our own custom good okay so let me quickly copy that code I have already written so this is the code actually uh this is these are the parameter you need to configure so I will just copy and I will paste it here okay so see these are the thing actually you need to mention so first of all you need to mention here plane dot txt file location okay so just copy the path and paste it here that's it okay Trend and inside validation just press the validation.txt this is the file and number of classes uh how many number of classes you have I have six number of classes that's why I have mentioned six and all the classes name you need to write here okay in the list okay so this is what actually you need to change inside this custom.yaml okay so once it is done you intercept okay you need to save this file and I will close this okay so yeah guys this was all about our data preparation okay for this model training so in our next video I'll be showing you like how to download a pretend model okay and uh how to start training process so yes guys thank you so much guys for watching this video and I will see you next time

Channel: DSwithBappy

Views: 6,075

Keywords: data science, machine learning, deep learning, python, computer vision, artificial intelligence, natural language processing, neural network, 100 days of ml, how to learn data science, dswithbappy, ds with bappy, bappy, 100 days of machine learning, মেশিন লার্নিং, ডাটা সায়েন্স, ডাটা সায়েন্স বাংলা, machine learning bangla, bangla ml tutorial, machine learning bangla course, data science bangla tutorial, data science bangla course, yolov7, object detection using yolov7

Id: bT05Iw2qyYA

Length: 5min 58sec (358 seconds)

Published: Wed Oct 19 2022