YOLOv8 in python environment for object detection | VSCode | OpenCV implementation of YOLO
Вставка
- Опубліковано 16 вер 2024
- The most recent and cutting-edge #YOLO model, #YoloV8, can be utilized for applications including object identification, image categorization, and instance segmentation. Ultralytics, who also produced the influential YOLOv5 model that defined the industry, developed YOLOv8. Compared to YOLOv5, YOLOv8 has a number of architectural updates and enhancements.
#computervision
#objectdetection
#ai
GitHub
github.com/DAV...
Man, THANK YOU SO MUCH! I was looking for this material for more than a week, and I found it from a blogger with 10 subscribers) it's just incredible!
You welcome 🙏
@@davidnyarko7300 hey..like i have train a seg model on my custom data this model is giving pretty good output on colab but when i am trying to test the same model VS code it is giving me problem...can you help me here ?
so far im in the minute 11 and it looks like the thing i am truly looking for. Thanks a lot man, love you so much, you're the only who teaches how to use the yolo model in python . PLease, let me spread your genes :D thanks
You are very much welcome.
Great content sir . Thank you for the efforts . As a request please continue making such informative and interesting content . 🙂
You are welcome
Explained really well. you have earned my subscription.
Thank you
THANK YOU VERY MUCH THIS VIDEO IS VERY HELPFUL
You are welcome
A great tutorial to get you started
Thank you
Thank you so much! I've been looking for weeks for more information on how to do this. everyone is using CLI instead
Glad it was helpful!
Nice Koby!!!!!
thank you so much, man. you helped me a lot :)
Amazing Koby! :)
can you please upload videos related to Object detection in REAL-TIME ASAP?
The Github page is updated with that, and you can use a WebCam as your input
Great video. You mentioned you will release similar video with training for custom class. Are you going to release it? I am waiting for it.
Sure, I am constantly going to be releasing from this week
Thanks!
Please comes out more videos like this!
Thanks, will do 🙏
good stufff mate thx
Thanks
@Koby_n_Code make some more videos bro, your content is exceptional! Can you make a video on implementing yolo into an arduino/raspberry pi?
Hi there. It shows the error called "TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first."
I had this too i fied it if you still care
you have to add .cpu() to the line
print(detection_output[0].numpy())
like print(detection_output[0].cpu().numpy())
Very good Video. Is it possible to train this Yolov8 with my custom dataset i created? I already have the images and the texfiles that contain the informations about the bounding boxes?
It is possible, I will be posting a video on that
@@davidnyarko7300 that would be great. Could you also explain how to use the gpu while training yolov8. I only manage to train on cpu which makes it almost impossible.
@@ranimaloui5931 Well noted
How did you set up the google collab storage session? I have been trying but I cant link my folders. I tried but I was unable to access the sub folders inside the folders. Each train, valid, and test folder contains two more folders called images and labels.
I believe the APIs have already changed since you made this video.
Could you share what version of YOLOv8 you are using?
The code is updated on the Github page i added
i have trained a yolov8-seg model on a custom dataset...but when i downloaded the model in my local system and tried it on vs code it is giving me unexpected output
thank you brooo
You are welcome
how can i make it so there is no output on the console whenever it detects the frames
You can comment out the print statement: If its coming from the the YOLO, you could try `contextmanager`
Hi there, thanks much. This is very useful. When is the next video coming up?
Thank you. I am starting a playlist for Object Detection and Segmentation this weekend.
very nice video but it will be better if you had a little better quality !!!
I found error when trying to make a rectangle with the same code , int() argument must be a string, a bytes-like object or a number, not 'Results'
Sentry is attempting to send 2 pending error messages
How can i solve it
I got the same error, have you solved it?
TypeError: int() argument must be a string, a bytes-like object or a number, not 'Results'
This issue has been solved as issue #2 on the Github page
It was associated with the new release of ultralytics and how they want to simplify the use of YOLO in a python environment.
get link bellow
github.com/DAVIDNYARKO123/yolov8-silva/issues/2
@@davidnyarko7300 @thank you very much!
@@andywong5193 You welcome
@@andywong5193 How did you modify the code to make it work?
Hi! We are creating a system that classifies tomato ripeness levels using image processing in CNN architecture with the YOLOv8 model. We are using Raspberry Pi 4 OS with 4GB RAM and we have encountered a problem - the system has 2-3 minute delay/lag in classifying the ripeness level. Would you happen to have any recommendation/suggestion sir on this problem?
Sure, my first recommendation to you is this video, you could get the Edge Tpu USB accelerator to help you speed up the process.
Very helpful video, great work. Is there any way I could do this with a url to a video stream?
Sure, just provide a UA-cam video link and it will work with it.
Hello good sir I know this is an old video but it is the one that has worked for me. Though I get an error: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first. DO you know of this if so thank you for any help you can give.
How do I make it so it does live image detection with my webcam?
Just replace the video link with `0` or any webcam index you have
Hye sir, i have run the yolo basic.py to test the image tensor, however the terminal shows None on dir. How to fix this? thankyou
What if we JUST want it to detect PERSON? And no other object?
Your class list should contain only "person"
Will detection work in grayscale images or videos?
hey brother no matter what i do its howing no module named 'torch' even though i am trying to install that it says no matching dist found for torch what should i do somebody help me
What system are you running on?
Hello, do u happen to know how to perform transfer learning and freezing?
Sure, but with which architecture are you looking at.
@@davidnyarko7300 like with the yolov8, is there like a certain command i just type into the command line so it executes freezing and like i can improve to my custom trained model?
What should I do if I want to create an output video at the end?
You will need to set “save” to True
github link is not working can u reply with the link ? thank you...!
Sure…
github.com/DAVIDNYARKO123/yolov8-silva
tysm for this tutorial! I just have a question, if I add a valid object that's not in the coco.txt file say "paper", can the model now detect paper?
No. you will need to train a new model on your custom dataset
Hey
I would like to know how we get the output on another terminal as you get as when I'm running the code it only giving the index and array as the output
You can set the show param to True.
is it possible if i use ESP32-cam as the live camera?
if i have a video that shows only a hand moving, how can i set a class hand to te model? The hand is classificated as 'person'
Currently, you can try the YOLO world model, you can set hand as a class.
how to count total object on 1 image detect?
You can get that from the length of the detections on each frame.
i have intel i5 and ram =8 G. but the video running slowly . and i have raspberry pi 3 cant running this program because the yolo algorithm need a more performance the CPU and RAM. Is my analysis correct?
Thank you very much for your excellent explanation.
Sure, other factors might cause your model to be slow. You can check my new video on how to get realtime speed on Raspberry Pi.
By the way, I live the way you teach, are you interested for particular classes? We can talk about the price
Sure! you can contact me at nyarkprince6@gmail.com
colab notebook mil sakti ha?
i need colab nootebook
What name command you write
Could you please let me know which specific command you are referring to? I'll be happy to help you with it.
Dear Koby, How could I use the GPU instead of the CPU?
Set device = [0], or device = “gpu”. Be sure you have GPU capacity on your system.
can you help me in detection with voicefeedback how i can do this? thanks waiting
Sure, but can you explain a bit further.
Source code!
Nice Tutorial. Can i get your email or linkedin account?
Prince David Nyarko is my LinkedIn name, it has same profile picture