Train Yolov10 object detection custom data FULL GUIDE | Computer vision tutorial

[@ComputerVisionEngineer]

Learn to create AI-based prototypes in the Computer Vision School! www.computervision.school 😃🚀🎓 ● Hands-On Computer Vision in the Cloud: Building an AWS-based Real Time Number Plate Recognition System bit.ly/3RXrE1Y ● End-To-End Computer Vision: Build and Deploy a Video Summarization API bit.ly/3tyQX0M ● Computer Vision on Edge: Real Time Number Plate Recognition on an Edge Device bit.ly/4dYodA7 ● Machine Learning Entrepreneur: How to start your entrepreneurial journey as a freelancer and content creator bit.ly/4bFLeaC

[@flourencelapore4404]

YOUR VIDS ARE REALLY HELPFUL TO US BEGINNERS! THANKS FOR THE AMAZING TUTORIAL!

[@ComputerVisionEngineer]

Thank you! 😃💪 Glad you enjoy the content! 🙌

[@ompatil3122]

Great stuff!!! Keep up the great content about Yolo models.

[@ComputerVisionEngineer]

Thank you for your support! I will keep up creating content about Yolo models! 😃🙌

[@cherrychingu]

Thank you for this amazing tutorial.

[@ComputerVisionEngineer]

You are welcome! 😃💪

[@HOOMANESMAILIAN-mf1yz]

hope you are doing well sir. why are you not energetic like before?? thanks for your amazing video.

[@ComputerVisionEngineer]

Thank you! I am still energetic! 😃 Just in a different way maybe 😄💪💪

[@sushanttwayana2302]

@ComputerVisionEngineer I really appreciate the content you’ve shared in this video. I encountered a specific error while running the code and was wondering if you could help me resolve it. Your guidance would be greatly appreciated. Thank you in advance!

[@giovanicostasilva941]

you are huge!! thanks so much

[@ComputerVisionEngineer]

Thank you! 😃🙌

[@fajrihanggi9759]

hello, its works, but how can i fix "no labels found in detect set, can not compute metrics without labels"?

[@talbiamal3282]

is there any platfom do the annotation fast for all images in the same time; or not; should we annotate the dataset manually?

[@bhukyapramod1923]

Can you say what could be the code/command line on colab if I wanted to test new image or video on model to predict?

[@coconutisland3572]

thanks alot, your video is very very useful ❣

[@ComputerVisionEngineer]

😄 You are welcome! Glad the content is helpful! 😃🙌

[@divye.ruhela]

1) How much VRAM is required to train it locally? 2) Is there a limit to how many classes you can train together? I may have to create 200+ classes to train some poses! 3) Does one have to go through drawing the annotations manually? As in, since I am going to train human poses, can't I have Yolo/ SAM create bounding boxes surrounding humans automatically and then use some tool to edit the labels on those boxes?

[@jonathanlew9706]

Appreciate the tutorial, but I keep getting the same error: 'DetMetrics' object has no attribute 'curves_results'. See valid attributes below. Any solutions?

[@flourencelapore4404]

Very life saver video sir, thank you sooo much. You just save my thesis project. But can i ask this one question, what if there are multiple images to classify? like a duck and a dog and a cat? how can i do that?

[@ComputerVisionEngineer]

Glad to hear the content is helpful! May I ask what your thesis is about? If you're referring to detecting multiple categories, the process is essentially the same. I’ll try to make another video on multi-class object detection in the future. 🙌

[@flourencelapore4404]

@@ComputerVisionEngineer I'm so glad to hear that good sir. By the way our thesis is about detecting/classifying the body fat of a person through object detection model. So we really need a multiple category. Looking forward to your coming vids sir. Thank you so muchhhh. You really saved our thesis

[@rataash]

Thank You.

[@ComputerVisionEngineer]

You are welcome! 😃🙌

[@cobweb6109]

You can solve Numpy Erro with this installment pip install albumentations==1.4

[@ronitroy8083]

thankyou bro

[@user-ic9yp6pe4j]

Hello! Thank you for this great tutorial. I was wondering in order to reproduce this, do I need to go through the annotation process in CVAT? Or would you be able to add all the datasets to the Github? Also, how can I run this on Google Collab with the environments you have? Would you be able to share a Google Collab link?

[@ComputerVisionEngineer]

Hello! I added the jupyter notebook I used in this tutorial to the project's github repository. Regarding the annotated data, you can download it from Google Open Images. 😃🙌

[@dynodyno6970]

Hello sir, I appreciate your video and guide. Could I follow these same steps for object detection within a video game? And just use screenshots and annotate for the game I want to detect in?

[@ComputerVisionEngineer]

Yes, you can apply the same process to any type of dataset, including videogame data. 🙌

[@cobweb6109]

First of all, thank you for sharing such a study. I ran the codes in google colab, but I got the error "TypeError: unhashable type: 'numpy.ndarray'" in ### 4. Train model ###. What should I do, can you help me?

[@SanilJadhav711]

Adding the !pip install albumentations==1.4 just before training should work for you In the last cell, just before you import Yolov10 from ultralytics, put the installation command, should look sumthn like this import os !pip install albumentations==1.4 from ultralytics import YOLOv10

[@udayannath8281]

​@@SanilJadhav711 Thanks a lot brother, it worked for me well.

[@SanilJadhav711]

@@udayannath8281 Glad to be of help to you

[@mujahidalisabir7689]

I love sir!! You are my crush in Computer vision. Are you available on LinkedIn?

[@ComputerVisionEngineer]

Thank you for your support! I am on LinkedIn, but not very active. 🙌

[@saraswati2143]

if i dont want to use the pre trained model that you are using in the video how can i train my model on my dataset ?? please reply i want to build a project on my own for college

[@ComputerVisionEngineer]

Are you asking if it's possible to train from scratch instead of using a pre-trained model? Yes, it's possible! Check the YOLOv10 GitHub repository's README for instructions. If I’m not mistaken, you just need to create the model with an empty constructor, like this: model = YOLOv10() 😃🙌

[@Baba-so6fh]

Is it possible to, intead of using testing this on a video, but rather to test this on a live video feed coming from a webcam?

[@ComputerVisionEngineer]

Yes, it is possible, you need to make only a few edits in the predict.py script. 😃🙌

[@mortysmith2192]

keep getting a type error, do you have any idea what could cause this ?

[@SanilJadhav711]

Adding the !pip install albumentations==1.4 just before training should work for you In the last cell, just before you import Yolov10 from ultralytics, put the installation command, should look sumthn like this import os !pip install albumentations==1.4 from ultralytics import YOLOv10

[@alvarorodriguez5395]

Hola, perdona que me haga un lío, pero podrías poner cual es la estructura de carpetas? Se que por un lado es images con las carpetas train y val en mi caso un 70% train 30% val y el la otra carpeta de label no me queda claro que va dentro de la carpeta de val. Veo que de pronto aparecen 500 labels. de donde han salido?

[@ComputerVisionEngineer]

Hola, la estructura sería así: - data --- images ------ train ------ val --- labels ------ train ------ val

[@alvarorodriguez5395]

@@ComputerVisionEngineer Vale, muchas gracias. Pero sigo con la duda de donde han salido los 500 archivos txt dentro de la carpeta de la labels

[@ComputerVisionEngineer]

@@alvarorodriguez5395 el dataset de imagenes y labels lo descargue de Google open images v7

[@viko4148]

Hi, could you show how to download from Open Imagen v7?

[@ComputerVisionEngineer]

Hi, I uploaded some simplified instructions on how to download data from Open Images v7 in my Patreon, available to my supporters here: bit.ly/46Ue4kC 🙌

[@RaptusAcht]

Thanks for the amazing video. But I can't train the model. Whenever I execute the fourth point in Google Colab, the program crashes because the RAM is completely used up. What can I do?

[@ComputerVisionEngineer]

How many images are you using? Are you using a session with gpu?

[@unknown_guy193]

Hello sir, I am getting every time typeError: unhashable type: 'numpy.ndarray' in google colab. I have checked the whole labels file and there is no error in the config.yaml. Can u help me on how to solve this?

[@chidieberemetu504]

Same issue with me

[@devikaskumar2207]

same issue, did you guys clear it?

[@devikaskumar2207]

heyy, did you guys tried running this is jupyter notebook?

[@halfwheeler]

@devikaskumar2207 @chidieberemetu504 i fixed this by running this in a cpu runtime and "!pip install albumentations==1.4"

[@flourencelapore4404]

me too this is the error that shows when i try this on google collab. Can someone help us?