YOLOv8: How to Train Objection Model with Custom Dataset

Inside my school and program, I teach you my system to become an AI engineer or freelancer. Life-time access, personal help by me and I will show you exactly how I went from below average student to making $250/hr. Join the High Earner AI Career Program here 👉 www.nicolai-nielsen.com/aicareer (PRICES WILL INCREASE SOON)
You will also get access to all the technical courses inside the program, also the ones I plan to make in the future! Check out the technical courses below 👇
_____________________________________________________________
In this video 📝 we are going to Train and Deploy a YOLOv8 for Custom Object Detection with Webcam in Real-time. First of all, we will train the yolov8 model in google colab on a custom dataset. After training, we will export the training YOLOv8 model and see how to deploy the model and run live inference on a webcam. This can actually run in real-time. You can also check out my YOLOv7 Course where we cover everything from generating a dataset, setting up the model, training, and deployment.
If you enjoyed this video, be sure to press the 👍 button so that I know what content you guys like to see.
_____________________________________________________________
🛠️ Freelance Work: www.nicolai-nielsen.com/nncode
_____________________________________________________________
💻💰🛠️ High Earner AI Career Program: www.nicolai-nielsen.com/aicareer
⚙️ Real-world AI Technical Courses: (www.nicos-school.com)
📗 OpenCV GPU in Python: www.nicos-school.com/p/opencv...
📕 YOLOv7 Object Detection: www.nicos-school.com/p/yolov7...
📒 Transformer & Segmentation: www.nicos-school.com/p/transf...
📙 YOLOv8 Object Tracking: www.nicos-school.com/p/yolov8...
📘 Research Paper Implementation: www.nicos-school.com/p/resear...
📔 CustomGPT: www.nicos-school.com/p/custom...
_____________________________________________________________
📞 Connect with Me:
🌳 linktr.ee/nicolainielsen
🌍 My Website: www.nicolai-nielsen.com/
🤖 GitHub: github.com/niconielsen32
👉 LinkedIn: / nicolaiai
🐦 X/Twitter: / nielsencv_ai
🌆 Instagram: / nicolaihoeirup
_____________________________________________________________
🎮 My Gear (Affiliate links):
💻 Laptop: amzn.to/49LJkTW
🖥️ Desktop PC:
NVIDIA RTX 4090 24GB: amzn.to/3Uc7yAM
Intel I9-14900K: amzn.to/3W4Z5Cb
Motherboard: amzn.to/4aR6wBC
32GB RAM: amzn.to/3Jt2XVR
🖥️ Monitor: amzn.to/4aLP8hh
🖱️ Mouse: amzn.to/3W501GH
⌨️ Keyboard: amzn.to/3xUGz5b
🎙️ Microphone: amzn.to/3w1F1WK
📷 Camera: amzn.to/4b4Ryr9
_____________________________________________________________
Timestamps:
0:00 Intro
0:39 Setup and Dataset
5:28 Train YOLOv8 Model
14:24 Deploying YOLOv8 Model
Tags:
#yolov8 #yolo #objectdetection #ultralytics

[@NicolaiAI]

Join My AI Career Program www.nicolai-nielsen.com/aicareer Enroll in My School and Technical Courses www.nicos-school.com

[@faisalhazry]

how do we want to make detection prediction classs 12:07

[@NicolaiAI]

@@faisalhazry i show that in another video on my Channel with yolov8 class

[@dakaradakaradakara699]

Sir please make a project tutorial like this, but with classifications custom dataset on roboflow.

[@HarmonyAham]

Hi. I was actually looking for the deployment process and skipped to that chapter in your video, but it looks like only inference is being shown. Where and how is it being deployed? Using kubernetes? TF serving?

[@dilaratuzuner3168]

you are always doing the best congratulations!!

[@NicolaiAI]

Thanks a lot!

[@harish3828]

one mistake or update in the yolo training line data={dataset.location} wont work so try source="dataset location"

[@tanhongshen6117]

thankssss , more people should see thiss

[@paulokafor3781]

this works

[@paulokafor3781]

data = "dataset location"

[@777sevenable]

thanks for that

[@neeraj.kumar.1]

Bro pls consider only one window of yourself because many information is getting hide.

[@kaka0199]

I have some problem with my virtual environment I installed all the required files and when I try to import YOLO from ultralytics for example from jupyter notepad using VScode, the import gets stuck and looks like it will never finish Maybe some of you have had a similar problem or know what is wrong with my environment?

[@8PassCharlie1]

hey can you tell is there anything else after import detection in line 2 of the code?

[@AdityaKumar-lv5st]

Where can I find the complete code for using the trained model for prediction? Its not completely visible in the tutorial video.

[@rachealcr6752]

Great video! Just wonder how to verify the evaluation results of mAP, Precision and recall in Roboflow using the results in Google colab?

[@OGIMxGaMeR]

Any idea on how to deploy this on a cloud server? How would you connect the cameras?

[@rakeshkumarrout2629]

This is really amazing. I really love your video. Kindly make a video on multiple video stiching and like making it live.

[@karimabdalla3136]

I get errors when passing to Install YOLO v8: ValueError: Invalid 'mode='. Valid modes are ('train', 'val', 'predict', 'export', 'track', 'benchmark').

[@user-vi2mp4vv4s]

Thankyou so much, Your video was short and simple and it helped me out a lot

[@NicolaiAI]

Glad it helped!

[@Zinou485]

my problem is stoping the session in kaggle or google colab and my volum dataset is 4005 image can you help me pleas Because I have my graduation thesis and I don't have much time

[@TheZombiebrainz]

I have trained how you do in the video but do not see how to export? can you help

[@kishorem5530]

sir if i place rtsp link instead of webcam source 0 , it freezes . Any solution for this sir.. Thank u

[@avionics412]

You videos are great ,I have been trying to count people at full screen or frames using yolov8 and with no success,what are the cods to do so I

[@YigalBZ]

Thanks for the good video. I will follow it. Can i get the coordinates of the detection objects real time? I would like to use them as an input for another program.

[@dgadstv3876]

may I ask sir, what code ediitor your are using?

[@fahmiraouin2983]

how can tunning hyperparameters for yolov8?

[@zenith0750]

How do i deploy it to mobile devices using flutter?

[@michaelaaron2134]

I’m constantly getting “FileNotFoundError”when training please what can i do

[@jabermobarak7]

how can I export the mode and get the path to it 14:30

[@syariffortask]

I using gtx1650,and why always using cpu intel core i5 gen 10 ram 16 gb process in my detection?how to optimize? Anybody can explain?

[@every-thing-is-ai]

tell me how to deploy yolov8 model on webcamp in colab

[@riaz_mahmud]

can you provide the colab notebook

[@Tetsujinfr]

It seems it does not like the reflectivity of the std cup at the end. Make sense, tricky case really, even for deepnet models. Thanks for the video, very cool nontheless. I like especially when you show stuff which do not work perfect on the webcam, that is interesting and genuine.

[@NicolaiAI]

Yeah for sure. Also the std cup was underrepresented in the dataset doe. I could have tuned some confidence score thresholds or added a tracker. It will make some false predictions when moving the camera around and when it runs with 75+ fps

[@princepatel1302]

im getting error while training the model

[@ahmedainouche5927]

Hi, Where is the code at 14:25 ?

[@bryanstanley4813]

@NicolaiAI thanks for the video! can I see the python code that you used in 14:25?

[@rajaaimannukmanrajazaharih8960]

hello sir i followed your tutorial and it worked out great! Do you have a video on how to extract the coordinates of the bounding box so i can determine the centroid of the boxes?

[@NicolaiAI]

Thanks a lot for watching! Yeah I have a video where we implement it in a custom class

[@jyleong2383]

**UPDATE** !yolo checks change content in -> data.yaml -> to train: ../train/images val: ../valid/images Image(filename=f'/content/runs/detect/train7/confusion_matrix.png', width=600)

[@petermaher7948]

you hero!

[@user-uq8pw4qs4k]

thanks bruh...

[@TimelessTales918]

An actual hero

[@limtless232]

Thanks a lot. In your code, line 5, how did you export and get the .pt file? I exported from roboflow as yolov8 but I don't get any pt file. thanks again :)

[@NicolaiAI]

U need to train the model as i do in this video. I then show how to download the model

[@xmarksmarko]

Hello, I love your videos and was wondering if you could make a video that covers in depth training process and datapreprocessing (such as: data augmentation, hyperparameter tuning, evolve, data vizualization etc.)

[@NicolaiAI]

Hi thanks a lot for watching! I already have videos going over all those things in my deep learning playlist

[@rakeshkumarrout2629]

Data Preprocessing is the most asked thing in this field I experienced this as an intern. I would request you to take the data and normalize and convert them to to nparrary Or do some Preprocessing. Kindly make it from a granular level

[@amrjawish8166]

thanks for the vid! can you make a video on how we can test our model for custom dataset (like a bunch of images on my local pc)?

[@NicolaiAI]

The first video I have about yolov8 I show how to use the command line for testing the Yolov8 models. I show with the webcam as the source but u can just throw in the path to the images or an image folder instead. Thanks for watching!

[@kelixoderamirez]

permission to learn and download dataset in roboflow sir. thank you

[@keshawnsmith320]

How did you get the results to save to 'runs/detect/predict/'? I have been struggling with this feat every time.

[@NicolaiAI]

It should do that automatically if u run the predict command

[@neverninetofive]

you have to add the argument save=True

[@ashishbhardwaj8609]

Hi, I was struggling with the python script to use it , so, thank you very much for uploading this tutorial! 😄 But, I still have one question, how can we deploy it using flask or some other framework?

[@AdityaKumar-lv5st]

HI, could you post the link to the python script for using the trained model for prediction. The complete code is not visible in video

[@gbo10001]

thanks for the video. how can i tackle a task of counting objects in a video with moving camera? specifically sea animals in a big pool so i need to scan all the pool and accumalte the numbers

[@NicolaiAI]

Thanks for watching! My tracking course with Yolov8 can exactly be used for that

[@gbo10001]

@@NicolaiAI t Does this course have support chat or something?

[@ahmetalperenaltundal9077]

hi, what webcam do you use in your projects

[@NicolaiAI]

I’m just using a cheap budget webcam

[@kin_1997]

awesome vidoes

[@NicolaiAI]

Thanks a lot! Really appreciate the support

[@houdabekkourialami3581]

Hello! I'm having trouble while trying to fix the epochs, i get an error that there are no images in the dataset although they are in fact existing in all of the train validation and test files. How can I fix ?

[@NicolaiAI]

There are some problems with the data.yml file. Replace the paths to the correct ones in there

[@houdabekkourialami3581]

Thank you so much. However I hvae another problem. When I run the vscode script I do get the results running in the terminal but the webcam isn't displayed. My webcam is activated and it does show up when I try functions like cv2.imshow()..

[@houdabekkourialami3581]

Never mind it works fine now. I had to update my environement.

[@abdulrasheedabdulmajeed5997]

Can you please do a tutorial video detecting potholes and crack using raspberry pi

[@gergessalahmorgan9182]

thank you very much

[@NicolaiAI]

Thanks for watching! Glad that it could help u

[@lamos1537]

Thanks for the video! Will you be making a video on training a custom dataset for segmentation on Yolo v8?

[@NicolaiAI]

I already did! Check out the second latest video

[@lamos1537]

@@NicolaiAI I saw thank you! What I mean is using your own dataset containing your own images. Do you maybe have a suitable tool that I can use to draw the bounding polygons for segmentation? I have used LabelImg already, but it obv is not going to cut it for instance segmentation datasets.

[@erikpetrosyan9662]

He has already made it!!! last video

[@lamos1537]

@@erikpetrosyan9662 that's my bad, I clearly didn't pay enough attention. Thanks

[@YigalBZ]

At 5:58 the "!yolo task=detect mode=train model=yolov8m.pt data={dataset.location}/data.yaml epochs=30 imgsz=640" command is not appearing on the notebook I copied. Should I add it manually? Also the 2 commands afterwards (start with image(filename)) are missing also from the notebook. If I run the following command (!yolo task=detect..) it fails. The error is: Traceback (most recent call last): File "/usr/local/bin/yolo", line 8, in sys.exit(entrypoint()) File "/usr/local/lib/python3.10/dist-packages/ultralytics/cfg/__init__.py", line 423, in entrypoint getattr(model, mode)(**overrides) # default args from model File "/usr/local/lib/python3.10/dist-packages/ultralytics/engine/model.py", line 372, in train self.trainer = trainer(overrides=overrides, _callbacks=self.callbacks) File "/usr/local/lib/python3.10/dist-packages/ultralytics/engine/trainer.py", line 127, in __init__ raise RuntimeError(emojis(f"Dataset '{clean_url(self.args.data)}' error ❌ {e}")) from e RuntimeError: Dataset '/content/Loco1-1/data.yaml' error ❌ Dataset '/content/Loco1-1/data.yaml' images not found ⚠, missing path '/content/Loco1-1/Loco1-1/valid/images' Note dataset download directory is '/content/datasets'. You can update this in '/root/.config/Ultralytics/settings.yaml'

[@tanhongshen6117]

same problem with me here

[@melkijonathan2475]

there a duplicate 'Loco1-1' here '/content/Loco1-1/Loco1-1/valid/images', you can delete the 'Loco1-1' in data.yaml train and val

[@houdabekkourialami3581]

Hello, where you able to fix it?

[@YigalBZ]

@@houdabekkourialami3581 No

[@kdlin1]

Where can I find the Colab scripts used in this KZbin video? Thanks.

[@NicolaiAI]

Thansk for watching! Think i deleted it when i updated the description. It is not added again. colab.research.google.com/drive/1q0GqWbYjACs9uVTjhsQfVQkfF0hhLHsq?usp=sharing

[@kdlin1]

@@NicolaiAI Thank you.

[@SaurabhSingh-mu5td]

Hii, very good work but how can i convert output in audio form so that it can be used by blind person

[@akhileshcruizes3637]

Try using pyttsx3 library

[@zellaphina5153]

Hi, you mention around 14:26 that you exported or downloaded the model so that you could use it with the webcam, but you don't describe that process at all. How do you get the pt file with your custom-trained model from Google Collab?

[@sevenplyy]

also still trying to figure out what happened there... solution?

[@Luca-yt2bg]

i am in the same question

[@Luca-yt2bg]

@@sevenplyy I FOUND

[@sevenplyy]

@@Luca-yt2bg NICE! please tell me

[@jabermobarak7]

@@Luca-yt2bghow?

[@sheicodes7393]

Superb informative! 🔥 May I ask for the Colab link? 🙏

[@NicolaiAI]

Thanks a lot! All the code is available on my GitHub

[@joart333]

Hello Brother I want the output results in json format do help me

[@NicolaiAI]

If u Watch my tracking video u Can see how to extract the results

[@joart333]

@@NicolaiAI Thanks for the reply brother

[@NicolaiAI]

@@joart333 of course!

[@divyeshjayswal3912]

hey, Can you provide me a script on how to deploy yolov8 model for intel Realsense Camera. Thanks in Advance! Btw your tutorials are very helpful. Keep up with the work!!

[@NicolaiAI]

Thanks a lot man! I already have a video where we implement a custom class with the yolo model and run inference on webcam then u will just have to swap the camera

[@panzach]

Thanks for this. However, the last part of the video isn't really "deploying" the model anywhere - it is about using it locally.

[@NicolaiAI]

Yes u can deploy models locally

[@muhammadasil9374]

@NicolaiAI Sir, please make a video on: how many sample or images of an object for training and then implementing, epoches, batch confidence value for achive better result. #nicolainielsen #nicolaiaI

[@NicolaiAI]

Thanks for the tip, will put that on my list of next videos to create. :-)

[@muhammadasil9374]

@@NicolaiAI thanks bro keep it up

[@NicolaiAI]

@muhammadasil9374 thanks bro!👊

[@tanhongshen6117]

For users who are unable to run the training process, be assured to change the data.yaml from the yolk file downloaded in the Collab. Go into the code, on the most bottom line. Change the file directory to ../xxxx/xxxx

[@NicolaiAI]

Thanks!

[@sherifkhaled1232]

can you show me the steps on how to do this please ?

[@houdabekkourialami3581]

How is that?