The models in this repository were trained using YOLO V11. The dataset consists of a collection of memes and publicly available images sourced from the internet.
This model is specifically optimized for online chat scenarios, ensuring that false positives (incorrect detections) are highly unlikely. However, false negatives (missed detections) may occur.
For instance, if a correctly identified "Nai Long" image achieves a similarity score between 0.9 and 1.0, a missed detection might produce a similarity score between 0 and 0.7 for the same image.
Validation Results:
For the validation dataset used in this model, the following metrics were recorded:
| Class | Images | Instances | Box(P) | R | mAP@50 | mAP@50-95 |
|---|
| all | 531 | 558 | 0.976 | 0.941 | 0.972 | 0.908 |
Prohibited Use:
The models in this repository are strictly prohibited from being used for any commercial purposes.
