This repository hosts a custom YOLOv8 object detection model trained to detect troop deployment boundaries in the game Clash of Clans. Specifically, it is trained to identify the valid deployable areas and boundary outlines for enemy bases that are Town Hall 11 and below.
The model was built using the Ultralytics YOLOv8 framework with the following configurations:
Below are the core indicators of how the model performed during the 300 epochs training! You can observe how the precision and recall stabilized as the dataset was digested.
To test and run this model locally using Python and the Ultralytics library, simply load the weights:
1from ultralytics import YOLO
2
3# Load the trained model
4model = YOLO('best.pt')
5
6# Perform predictions on a Clash of Clans screenshot
7results = model('path/to/your/screenshot.jpg')
8
9# Show and save results
10results[0].show()
