metadata
datasets:
- pierreguillou/DocLayNet-large
Ultralytics YOLOv8 small trained on DocLayNet-large. Training hyperparameters can be found on the args.yaml file.
Val
| Metric | Value |
|---|---|
| box_loss | 0.71296 |
| cls_loss | 0.51492 |
| dfl_loss | 0.81419 |
| precision(B) | 0.891 |
| recall(B) | 0.8338 |
| mAP50(B) | 0.909 |
| mAP50-95(B) | 0.711 |
Test
| Metric | Value |
|---|---|
| Precision | 0.852 |
| Recall | 0.822 |
| mAP50 | 0.891 |
| mAP50-95 | 0.733 |