How to Use YOLOv8 Python for Computer Vision Tasks

Use the ultralytics Python package to load a YOLOv8 model with YOLO('yolov8n.pt'), then run detection on images or videos using model.predict(). This lets you quickly detect objects in computer vision tasks with minimal code.

Syntax

The basic syntax to use YOLOv8 in Python involves importing the YOLO class from the ultralytics package, loading a pre-trained model, and running predictions on images or videos.

• YOLO('model_path'): Loads the YOLOv8 model from a file or pretrained weights.
• model.predict(source='image_or_video_path'): Runs object detection on the given source.
• results: Contains detected objects, bounding boxes, and confidence scores.

python

from ultralytics import YOLO

# Load a pretrained YOLOv8 model
model = YOLO('yolov8n.pt')

# Run detection on an image
results = model.predict(source='image.jpg')

# Access detected boxes and classes
boxes = results[0].boxes
classes = results[0].boxes.cls

Example

This example shows how to load the YOLOv8 nano model, run detection on a sample image, and print detected object classes and confidence scores.

python

from ultralytics import YOLO

# Load YOLOv8 nano model
model = YOLO('yolov8n.pt')

# Run prediction on an example image
results = model.predict(source='https://ultralytics.com/images/bus.jpg')

# Print detected classes and confidence scores
for result in results:
    for box in result.boxes:
        cls = model.names[int(box.cls)]
        conf = box.conf.item()
        print(f'Detected {cls} with confidence {conf:.2f}')

Output

Detected bus with confidence 0.98
Detected person with confidence 0.85
Detected car with confidence 0.76

Common Pitfalls

• Not installing the ultralytics package first with pip install ultralytics.
• Using incorrect model file names or paths causes loading errors.
• Passing invalid source paths to model.predict() leads to file not found errors.
• Forgetting to convert detected class indices to names using model.names.
• Not handling results as a list of detections; always index the first element for single image input.

python

from ultralytics import YOLO

# Wrong: missing model file or wrong name
# model = YOLO('wrong_model.pt')  # This will raise an error

# Correct:
model = YOLO('yolov8n.pt')

# Wrong: invalid source path
# results = model.predict(source='nonexistent.jpg')  # FileNotFoundError

# Correct:
results = model.predict(source='image.jpg')

Quick Reference

Here is a quick summary of key YOLOv8 Python usage tips:

• Install with pip install ultralytics.
• Load model: model = YOLO('yolov8n.pt').
• Run detection: results = model.predict(source='image.jpg').
• Access detections: results[0].boxes for bounding boxes.
• Get class names: model.names[int(box.cls)].
• Supports images, videos, webcam streams as source.

Key Takeaways

Install the ultralytics package to use YOLOv8 in Python easily.

Load pretrained models with YOLO('model_name.pt') and run detection with model.predict().

Always check that source paths and model files exist to avoid errors.

Use results[0].boxes and model.names to interpret detection outputs.

YOLOv8 supports images, videos, and live streams for flexible computer vision tasks.