TRAIN YOLO-V5 ON CUSTOM DATASET || YOLO-V5 - FACE MASK DETECTION

[@shakilahmad2282]

I watched a lot of videos over the last week. Everyone would start from scratch, but by the middle, they would have left the practical work unfinished. Just got a video that worked very well until the end. Thank you for such amazing video.

[@coder_zero]

😃😃😃

[@yasminbanu2597]

Hello sir have implemented this with 5 classes and got good results. Thank you for your video

[@coder_zero]

I would suggest to get more data and train the model for more number of epochs😀😀😀

[@515rishitha4]

what are those 5 classes may i know...?

[@syukrikoko5975]

Thank you for the video and knowledge... I'm looking forward to your next video.

[@coder_zero]

Thank you 😃

[@MushasFACT]

you are god bro thanks a lot

[@coder_zero]

You are so kind 😀

[@gzhrk]

thank you so much

[@coder_zero]

😃

[@itspiyu2536]

Kya aap no plate detection keliye ye train model ka alag video dal sakte ho ?? Plz using video file only

[@coder_zero]

Hi, no need for that. Just replace the data and dataset.yaml file with your data and yaml file accordingly. Follow the tutorial. You will have your model :)

[@itspiyu2536]

Sir but where is dataset

[@streamovie5601]

Just amazing , waiting for more videos please, thanks man

[@coder_zero]

Thanks, will do!😄

[@merondesta3895]

thankyou sir that was useful. i trained using my own dataset but no matter how long i trained it the m@p will not get improved from 0.34 something like that i have used same data for yolov4 but it performs well

[@coder_zero]

Hi, data annotations for all yolo models will be same. To get better results, we have to create a good distribution of data ( day and night time , different perspectives, different lighting conditions, etc .) 😃

[@kuldeepaher4937]

Thank you very much bro..... keep uploading such videos

[@coder_zero]

:) :)

[@shalinikumar5621]

Detailed explanation 👍

[@coder_zero]

Thank you :)

[@gzhrk]

Hi i have question please , what software you use to labeling yiur image , is my first time i will try to do this like your video , if also can you put the code to train the image ???

[@coder_zero]

LabelImg

[@gzhrk]

@@coder_zero thank you so much , i will try to do like video to train my images

[@RAZZKIRAN]

amazing , thank you.

[@coder_zero]

😃

[@Dr.AhmedQusaySabri]

thank you

[@coder_zero]

😃

[@abdalnaeem5560]

useful content

[@coder_zero]

Thank you😃

[@You2ubyTunes]

got everything down to uploading the files to google now am lost. do i have to make a notebook ? how do i do that ?

[@sing_with_gs]

How to use this on live camera please give guidance sir

[@coder_zero]

Hi please have a look at this video : kzbin.info/www/bejne/n6alqqipbLh2e6M

[@youness4295]

Thanks for sharing

[@coder_zero]

♥️

[@JohnJoe131]

Yo bro how would you go about changing this to a real time set up, as in with a webcam

[@coder_zero]

Hi, please have a look at this video : kzbin.info/www/bejne/n6alqqipbLh2e6M

[@Alpha-hj2ss]

Thank you. Great Tutorial

[@coder_zero]

Thank you 😃

[@ajitegar7738]

THANKS ! 🔥

[@coder_zero]

😀😀

[@techease207]

how can i make a flask webapp using this model can you assist me with that?

[@RAHUDAS]

Is there any keras implementation of the same

[@coder_zero]

I think for now, there is no keras implementation. But, you can convert into Tensor RT for faster inference.

[@itspiyu2536]

Number plate detection ke liye bhi same Krna hai??

[@coder_zero]

Yup, just change the class name accordingly

[@taeefnajib]

I have a question. How can I save and download the model and re-use it in my local machine without training? I'd like to test it using my webcam live stream, which is not possible on Colab. But I don't have a GPU, so I need to train the model on Google Colab for the higher computation power and reuse the model in my local machine. Is that possible?

[@coder_zero]

Hi, yes it is possible. In the video I have given information on how to download the trained model and please watch this video kzbin.info/www/bejne/n6alqqipbLh2e6M for inference on local machine

[@coder_zero]

Also when you are loading the model on local machine using pytorch hub set device = 'cpu' if you have no gpu

[@shounakkshirsagar2009]

1)Is it necessary to clone the Yolov5 git repo in the same drive and folder where we save our train/test images? 2)I have cloned the yolov5 git repo in C drive [C/yol5/yolov5] and my train/test images are in E drive under img_data folder (Train= E/img_data/train Test= E/img_data/test) Here, how should I specify the path in dataset.yaml for train: and test: P.S :I’m using anaconda prompt for running the training command for yolov5

[@coder_zero]

It is not necessary to keep images and git repo at same location. You should be very careful with the train and test folder paths while creating the dataset yaml file.

[@amirulirfanbinnazaruddin7317]

can you teach me how to put live webcam in it

[@alexkulikov5673]

Try this: sourse 0

[@coder_zero]

Hi, please have a look at this video : kzbin.info/www/bejne/n6alqqipbLh2e6M

[@taeefnajib]

@@alexkulikov5673 You can't use source 0 on colab. However you can do that on your local machine.

[@saipavan5194]

Hi i am predicting credit card start date and end date in my project using yolov5 model.. can you please help me on how to get only end date as an output

[@coder_zero]

Hi, I am not getting a clear idea about the project and problem. Can you please elaborate?

[@motoraddiction6412]

If I train yolov5 for license plate detection, are there any changes to do ACC. to this video? Especially in dataset.yaml.

[@coder_zero]

Yeah. You need to change the class name. Also please go through the tutorial again for better understanding 😃

[@duseokchoi2312]

Inside train file, order of images and labels have to be same or doesnt matter? for etc, image 1 , image 2, image 3 label of image 3 , label of image 2 , label of image 1

[@coder_zero]

Order does not matter. Annotation file for an image should be present in the label folder.

[@duseokchoi2312]

@@coder_zero one more question, I am having really weird issue that my image 1 and image 2 have different label, but only one of them get trained no matter what although they are both in same images and label file. What problem might causing this issue?

[@gauthamkrishna2660]

can u share the full code line of "inference or detection on new images"

[@gauthamkrishna2660]

that code line is not completely written in ur github also..pls reply

[@coder_zero]

Please have a look at the notebook. The code is there.

[@Happykittooo]

From where can i find test folder ? i've downloaded dataset, but there is only train and val folder, but no test folder .. so where can i find it ?

[@coder_zero]

Hi, you can download images from Google to test your model. In this way you can find out your model performance 😁

[@KYPUNAGENDRAREDDY2021-CSEUG]

bro how to find its accuracy

[@coder_zero]

Accuracy is not the best evaluation parameter for an object detection model. mAP should be used to evaluate an object detection model.

[@515rishitha4]

sir i am getting an error while i am retraining the model from last weights

[@coder_zero]

What is the error?

[@justdoit35]

how to find accuracy of model?

[@coder_zero]

Hi, for object detection models we need to focus on the mAP and losses rather than accuracy.

[@justdoit35]

@@coder_zero thanks ! how to implement it into yolo V3?

[@FCL00]

can i use yolov5 for face detection?

[@coder_zero]

Yes, you can use yolov5 for face detection.

[@umar_muhammad_yaree]

Hello sir nice tutorial How do I count number of objects in a picture or video?

[@coder_zero]

To create an effective counting algorithm, you have to track each and every objects by assigning unique IDs and count these objects when they cross a certain ROI. Tracking is necessary for removing duplicate counts i.e. counting the same object multiple times.

[@umar_muhammad_yaree]

@@coder_zero thank you for the reply And please how do I do that

[@crikshikshow9189]

I am getting an error after following all your steps can you help me?

[@coder_zero]

What kind of error ?