joaolsilva
commited on
Commit
•
f58cc92
1
Parent(s):
f68b056
Update README.md
Browse files
README.md
CHANGED
@@ -1,4 +1,27 @@
|
|
1 |
---
|
2 |
datasets:
|
3 |
-
-
|
4 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
datasets:
|
3 |
+
- pierreguillou/DocLayNet-large
|
4 |
+
---
|
5 |
+
|
6 |
+
Ultralytics YOLOv8 nano trained on DocLayNet-large. Training hyperparameters can be found on the args.yaml file.
|
7 |
+
|
8 |
+
### Val
|
9 |
+
|
10 |
+
| Metric | Value |
|
11 |
+
| --- | --- |
|
12 |
+
| box_loss | 0.74732 |
|
13 |
+
| cls_loss | 0.52909 |
|
14 |
+
| dfl_loss | 0.81974 |
|
15 |
+
| precision(B) | 0.87111 |
|
16 |
+
| recall(B) | 0.81897 |
|
17 |
+
| mAP50(B) | 0.89472 |
|
18 |
+
| mAP50-95(B) | 0.6866 |
|
19 |
+
|
20 |
+
### Test
|
21 |
+
|
22 |
+
| Metric | Value |
|
23 |
+
| --- | --- |
|
24 |
+
| Precision | 0.851 |
|
25 |
+
| Recall | 0.798 |
|
26 |
+
| mAP50 | 0.879 |
|
27 |
+
| mAP50-95 | 0.712 |
|