Train Yolov9 object detection custom data on Google Colab | Computer vision tutorial

Code: github.com/computervisioneng/...
🎬 Timestamps ⏱️
0:00 Intro
0:22 Yolov9 repository (fork)
2:52 Google colab notebook
4:22 Data
13:29 Model training
16:15 Get results
19:20 Alternative to Google Colab
21:04 Outro
🌍 Community 👥
Join our Discord server: / discord
Support me on Patreon: / computervisionengineer
#computervision #python #yolov9 #yolo #objectdetection

[@ComputerVisionEngineer]

Did you enjoy this video? Try my premium courses! 😃🙌😊 ● End-To-End Computer Vision: Build and Deploy a Video Summarization API bit.ly/3tyQX0M ● Hands-On Computer Vision in the Cloud: Building an AWS-based Real Time Number Plate Recognition System bit.ly/3RXrE1Y ● Machine Learning Entrepreneur: How to start your entrepreneurial journey as a freelancer and content creator bit.ly/4bFLeaC All my premium courses are available to the Computer Vision Experts in my Patreon. 😉 www.patreon.com/ComputerVisionEngineer

[@noobtopro8699]

how to do detection using your code, how to pass the trained model to detect

[@EtienneDescombes]

Very useful thx :)

[@gaeunjung7476]

There is a small error. path: /content/data/ => path: /content Thanks for sharing your knowledge😊

[@user-ky1ug2ex5x]

how can we get the accuracy , precision and other things please provide link or blog or include in video

[@Kaori-ei1gt]

Hi I would like to ask how do we train other YOLOV9 models?

[@selfimprovement5249]

i have best.pt and last.pt in weights folder and i just want to do inference on the input image help me

[@user-xw7ks6qg2q]

which annotation tool do recommend for a custom dataset? thanks a lot for the video, very cool very interesting!!

[@ComputerVisionEngineer]

Glad you enjoyed it! I usually use cvat. 🙌

[@abrarahmed3495]

Hi, Just a question here, I wanted to run the model on the test images to see if the model works or not, do you have the code for that?

[@sahilshirodkar6040]

how do i make real time detections using this? Thanks in advance

[@bikalbaral1394]

Hello, so basically, if we get this file "file.pt" , we can detect objects by using the code right beacuse file.pt will contain all the trained information there, can you please provide one .pt file for me , i would be happy for that ,regards! I am trying to detect objects for , I already trained the model, got the pt file but is is saying ,no any objects detected. Please ,help me!

[@chiragmetgud4270]

Sir can you tell whether you are using the yolov9e.pt weights file for training as its more efficient, if not how to do the same according to your code format, as I need accurate and correct predictions.

[@ComputerVisionEngineer]

I am using yolov9c.pt, please take a look at the wrapper I created in yolov9.py to see how to change the code to use yolov9e.pt instead.

[@selfimprovement5249]

I want to do inference on input image please help me . using yolov9 weights file

[@blackboltmedia6822]

Sir, how to install uints I tried but I don't know how to install in vscode

[@dkryptonut]

Felipe, do you know if this will work with the YOLOv9 MIT rewrite (made by the same group)? I think a few people would be interested as it has a more permissive license than GPLv3.

[@ComputerVisionEngineer]

Good question, I will take a look.

[@nguyenha6635]

Sir i dont know why i did all step but my yolov9 cant resolved to train

[@uniqx-ai]

Is it possible to detect ocean waves, track them, and measure their height as well? If so, are the methods the same?

[@ramazan1250]

Detection and tracing may be possible. But measuring heights is very complicated, even impossible. And i think yolov8 is easier if you want to detect waves.

[@ComputerVisionEngineer]

Yes, it may be challenging to measure the heights, it depends on your data.

[@uniqx-ai]

I am curious about the use of tasks yolov8/9 for object measurement. Does it use segmentation? If there is a video tutorial for object measurement complete with how to prepare the dataset, it would be interesting. If the speed of an object can be measured, then the size of an object should also be measurable, shouldn't it?@@ComputerVisionEngineer

[@videospool2342]

Can you please share these files along with yolov9 repository on the KZbin channel for free download purpose....

[@ComputerVisionEngineer]

Hi, the github repository is linked in the description 🙌

[@karimmerchaoui9736]

I couldn't find the video about downloading a dataset from open images dataset

[@ComputerVisionEngineer]

Hi, the video is available in my Patreon.

[@ComputerVisionEngineer]

Here: bit.ly/46Ue4kC

[@lethnis9307]

Could you use pretrained model?

[@ComputerVisionEngineer]

Doing transfer learning from a pre trained model should be possible, I am going to look into it.

[@Mrsmith0119]

Can you tell us how to increase the accuracy of my model

[@ComputerVisionEngineer]

There are many ways to increase the accuracy of your model, tell me more about your project; how many classes are you trying to detect and how many images do you have?

[@vishnuvardhannanda8606]

@@ComputerVisionEngineer I have 620 original images augmented to 1400 images ( 1200-100-100 for train val test) for 3 classes. my project is to detect dents/damages/scratches on motorbikes. give your suggestions pls how you gonna do which models/weights you'll use

[@ramyeonmokja]

@@vishnuvardhannanda8606 heyy! how did you increase your accuracy?? I am doing a similar project- conveyor belt tear detection and my dataset size is also same. I'm really struggling and could use any tips or help you can offer!

[@jimjambles8918]

NOTE TO VIEWER... IF YOU ARE NOT A PATREON FOLLOWER OF THIS GUY, DON'T WASTE YOUR TIME WATCHING THIS VIDEO. NO LINK TO TRAINING DATA + PUSH FOR PATREON IN ORDER TO GET SAMPLE DATA. I DON'T LIKE TUBERS WHO UPSELL WITH ONLY HALF USABLE PROCESS.

[@ComputerVisionEngineer]

Thank you for your feedback! Promoting my Patreon helps cover some expenses and allows me to continue creating content on KZbin. Additionally, if you want to download data from the Google Open Images v7, you can follow the instructions provided on their website. The content I share on my Patreon is aimed at simplifying the download process.

[@jimjambles8918]

@@ComputerVisionEngineer cool story bro. It's a sales funnel. Just like every other youtuber.