metadata
datasets:
- pierreguillou/DocLayNet-large
Ultralytics YOLOv8 nano trained on DocLayNet-large. Training hyperparameters can be found on the args.yaml file.
Val
| Metric | Value |
|---|---|
| box_loss | 0.74732 |
| cls_loss | 0.52909 |
| dfl_loss | 0.81974 |
| precision(B) | 0.87111 |
| recall(B) | 0.81897 |
| mAP50(B) | 0.89472 |
| mAP50-95(B) | 0.6866 |
Test
| Metric | Value |
|---|---|
| Precision | 0.851 |
| Recall | 0.798 |
| mAP50 | 0.879 |
| mAP50-95 | 0.712 |