Update README.md
Browse files
README.md
CHANGED
|
@@ -54,7 +54,7 @@ Test labels are hidden to ensure fair evaluation.
|
|
| 54 |
The baseline model is a **YOLOv12** multitask variant, extended with a **regression head** for predicting food weight (see Figure below). It was designed to be **agnostic to missing labels**, making it compatible with datasets that do not have weight annotations.
|
| 55 |
|
| 56 |
|
| 57 |
-
|
| 58 |
|
| 59 |
### Best Model (YOLOv12-M @ 640×640):
|
| 60 |
- **Detection**: mAP50 = 0.974, mAP50-95 = 0.948
|
|
|
|
| 54 |
The baseline model is a **YOLOv12** multitask variant, extended with a **regression head** for predicting food weight (see Figure below). It was designed to be **agnostic to missing labels**, making it compatible with datasets that do not have weight annotations.
|
| 55 |
|
| 56 |
|
| 57 |
+
Github Source Code: [Multitask-Food-Portion-Estimation](https://github.com/IS2AI/Multitask-Food-Portion-Estimation)
|
| 58 |
|
| 59 |
### Best Model (YOLOv12-M @ 640×640):
|
| 60 |
- **Detection**: mAP50 = 0.974, mAP50-95 = 0.948
|