Train YOLOv8 Classification on Your Custom Dataset | Step By Step Guide

In this in-depth tutorial, we'll guide you through the process of training YOLOv8 classification models on your very own custom dataset. YOLOv8 is a state-of-the-art algorithm for object detection and classification, and in this video, you'll learn how to leverage its capabilities to classify objects in your unique dataset.
🔥 Step-by-Step Guide 🔥
From installing YOLOv8 to exporting the trained model, we cover each step comprehensively. Follow along to master the essential techniques for building your classification model.
⭐️ What You'll Learn ⭐️
Verify GPU Access: Make sure you have access to a GPU for faster processing.
Installation: Install YOLOv8 using the pip package and validate the installation.
Dataset Preparation: Organize your custom dataset with the correct folder structure.
Download the Characters Dataset: Use a pre-existing dataset for demonstration purposes.
Custom Training: Dive into the world of YOLOv8 classification training with your own dataset.
Validate Your Model: Evaluate the performance of your trained model on a validation dataset.
Inference: Witness the magic as the custom model classifies objects in new images.
Export Your Model: Learn how to export your YOLOv8 model to different formats for deployment.
Download Your Trained Model: Get your hands on the final trained model to use in your projects.
📂 Get the Notebook and Dataset 📂
Access the notebook code and characters dataset link in the pinned comment or video description.
🔔 Subscribe for More AI Content 🔔
Don't forget to subscribe to our channel and hit the notification bell so you won't miss any future tutorials and updates.
👍 Like and Share 👍
If you find this tutorial helpful, give it a thumbs up and share it with others who might be interested in learning YOLOv8 classification.
Notebook and Dataset: drive.google.com/file/d/1caME...
YOLOv8 Official Website: ultralytics.com/yolov8
Ultralytics GitHub: github.com/ultralytics/ultral...
Get ready to embark on an exciting journey of YOLOv8 classification training! Let's dive in and build your custom model for object classification on your dataset. 💻🚀
#YOLOv8 #ObjectClassification #CustomDataset #AI #Tutorial #yolov8 #artificialintelligence #artificiallyIntelligent #custom
#training

[@viewview6687]

thank you sir for the great and easy to understand video.

[@ArtificiallyIntelligents]

you're welcome

[@headwellledesma7793]

What if my dataset is in my google drive..

[@MALLASWETHA]

In results.csv file there js not gt and pred columns?

[@theinfamous4231]

how can I download YOLOV8 module locally so that I can train my custom data locally

[@MALLASWETHA]

How can we find out the accuracy , recall curves????

[@DhanushPalla]

sir how to find precision f1score recall precvision in yolov8 classification

[@ArtificiallyIntelligents]

You can see docs.ultralytics.com/reference/utils/metrics/#ultralytics.utils.metrics.ConfusionMatrix.tp_fp for finding the precision, recall, F1 Score for classification model

[@yashasvikedia5889]

Sir we did the training following your video above. We are unable to test the model by uploading a sample image. Can you please help us with that?

[@moazzam99]

Can you please tell me about the error exactly that you are getting

[@danishzia8192]

Please tell me we don't need labels for classification in yolov8?

[@inckaje]

yes you don't need it

[@ahmedprog3174]

Thank you very
Can you help me for print fscor ,precition in this code

[@ArtificiallyIntelligents]

Yeah sure
After training the mode, you will have a results csv file into your train directory where all the training results are stored.
# Load the validation results CSV file
validation_results = pd.read_csv(f'{HOME}/runs/classify/train/results.csv')
# Extract ground truth and predicted labels
true_labels = validation_results['gt'].tolist()
predicted_labels = validation_results['pred'].tolist()
# Calculate F1 score and precision
f1 = f1_score(true_labels, predicted_labels, average='weighted')
precision = precision_score(true_labels, predicted_labels, average='weighted')
# Print the calculated metrics
print(f'Weighted F1 Score: {f1:.4f}')
print(f'Weighted Precision: {precision:.4f}')

[@ahmedprog3174]

Thank you for your reply. The explanation is really clear. I wish you a good luck❤

[@ahmedprog3174]

Sorry i have one problem [gt] [pred]
How to import?

[@DhanushPalla]

sir how to find precision f1score recall precvision in yolov8 classification@@ArtificiallyIntelligents

[@daryljaysagabaen9562]

TypeError: expected str, bytes or os.PathLike object, not NoneType
error appear when I validate the model

[@ArtificiallyIntelligents]

Please verify that the paths are correct according to your drive paths

[@MALLASWETHA]

Is data.yaml not necessary For training the yolov8?????

[@55_it_kuldeepkumarsaini26]

yes ! I also getting the same error .. that data.yaml not found

[@prasannaravijr1420]

CAN YOU PLS HELP ME SOLVE THIS ERROR:
"RuntimeError: Trying to resize storage that is not resizable"

[@ArtificiallyIntelligents]

Can you please tell me where you are getting this error

[@prasannaravijr1420]

@@ArtificiallyIntelligents !yolo task=classify mode=val model={HOME}/runs/classify/train/weights/best.pt data='{DATA_DIR}'

[@nazilaameli8639]

@@ArtificiallyIntelligents Hello, and thanks for the wonderful video. I have the same problem after trying to run the code %cd {HOME}
!yolo task=classify mode=val model={HOME}/runs/classify/train3/weights/best.pt data='{DATA_DIR}'.

[@prasannaravijr1420]

@@nazilaameli8639 @ArtificiallyIntelligents If you guys found any solution pls let me know

[@boringtaskai]

@@nazilaameli8639 Hi, as describe in ultralytics GitHub repositories issues, itt is a bug from yolo version 8.0.231, the temporary solution is to downgrade the yolo to 8.0.180
So you can change the part which install the yolo to use 8.0.180 version instead.
!pip install ultralytics==8.0.180