Upload mira_exp018.yaml with huggingface_hub
Browse files- mira_exp018.yaml +20 -0
mira_exp018.yaml
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
name: EXP-018 YOLO11n Teacher
|
| 2 |
+
display_name: MIRA EXP-018 YOLO11n Teacher
|
| 3 |
+
class_names:
|
| 4 |
+
- glass
|
| 5 |
+
- metal
|
| 6 |
+
- paper
|
| 7 |
+
- plastic
|
| 8 |
+
- trash
|
| 9 |
+
model_type: yolo_pt
|
| 10 |
+
model_file: mira_exp018.pt
|
| 11 |
+
metrics:
|
| 12 |
+
mAP50: 0.906
|
| 13 |
+
mAP50_95: 0.822
|
| 14 |
+
tags:
|
| 15 |
+
- teacher
|
| 16 |
+
- balanced
|
| 17 |
+
- no_sortwaste
|
| 18 |
+
- kaggle_t4
|
| 19 |
+
- adamw
|
| 20 |
+
description: YOLO11n teacher trained on balanced 4-source dataset (dmedhi+TACO+Roboflow+TrashNet, 5108 train/415 val). Achieves 90.6% mAP50 on tabletop TrashNet validation. INT8 TFLite export 2.90 MB.
|