[FEATURE] Model (best.pt and last.pt) and model card upload for yolov8n celebrity detection model
Browse files- README.md +99 -1
- images/.gitattributes +1 -0
- images/F1_curve.png +3 -0
- images/P_curve.png +3 -0
- images/R_curve.png +3 -0
- images/confusion_matrix_normalized.png +3 -0
- images/image1.png +3 -0
- images/image2.png +3 -0
- images/image3.png +3 -0
- images/results.png +3 -0
- models/.gitattributes +1 -0
- models/best.pt +3 -0
- models/last.pt +3 -0
README.md
CHANGED
@@ -1,3 +1,101 @@
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Celebrity Detection YOLOv8n
|
2 |
+
|
3 |
+
📌 **[English](#celebrity-detection-yolov8n) | [Español](#espanol)**
|
4 |
+
|
5 |
---
|
6 |
+
|
7 |
+
## 📄 Celebrity Detection YOLOv8n (English)
|
8 |
+
⬇ Scroll down for [Español](#espanol)
|
9 |
+
|
10 |
+
This repository contains a **YOLOv8n** model trained to detect the main characters from the TV show **Friends** in images and videos. The model was trained on a custom dataset labeled using Amazon Rekognition.
|
11 |
+
|
12 |
+
## 📌 Model Details
|
13 |
+
|
14 |
+
- **Model:** YOLOv8n (pretrained on COCO, fine-tuned on Friends dataset)
|
15 |
+
- **Classes:**
|
16 |
+
- Matt LeBlanc (Joey)
|
17 |
+
- David Schwimmer (Ross)
|
18 |
+
- Matthew Perry (Chandler)
|
19 |
+
- Courteney Cox (Monica)
|
20 |
+
- Jennifer Aniston (Rachel)
|
21 |
+
- **Dataset:** Custom dataset generated from **YouTube shorts** of Friends episodes
|
22 |
+
- **Training Images:** ~13,000 (80% training, 20% validation)
|
23 |
+
- **Hardware Used:** Trained on GPU
|
24 |
+
- **Frameworks & Libraries:** `ultralytics`
|
25 |
+
|
26 |
+
## 🏆 Model Performance
|
27 |
+
|
28 |
+
| Metric | Value |
|
29 |
+
|------------------|-------|
|
30 |
+
| **mAP@50** | 0.98 |
|
31 |
+
| **mAP@50-95** | 0.91 |
|
32 |
+
| **Precision** | 0.96 |
|
33 |
+
| **Recall** | 0.98 |
|
34 |
+
|
35 |
+
### Performance Curves:
|
36 |
+
|
37 |
+
**F1-Confidence Curve:**
|
38 |
+

|
39 |
+
|
40 |
+
**Recall-Confidence Curve:**
|
41 |
+

|
42 |
+
|
43 |
+
**Precision-Confidence Curve:**
|
44 |
+

|
45 |
+
|
46 |
+
**Confusion Matrix:**
|
47 |
+

|
48 |
+
|
49 |
+
## 📸 Example Detections
|
50 |
+
|
51 |
+
Correctly detected images:
|
52 |
+

|
53 |
+

|
54 |
+
|
55 |
+
Incorrect/misclassified example (blonde woman mistaken for Jennifer Aniston):
|
56 |
+

|
57 |
+
|
58 |
---
|
59 |
+
|
60 |
+
<a id="espanol"></a>
|
61 |
+
## 📄 Modelo de Detección de Celebridades YOLOv8n (Español)
|
62 |
+
⬆ Scroll up for [English](#celebrity-detection-yolov8n)
|
63 |
+
|
64 |
+
Este repositorio contiene un modelo **YOLOv8n** entrenado para detectar a los personajes principales de la serie **Friends** en imágenes y videos. El modelo fue entrenado en un conjunto de datos personalizado etiquetado con Amazon Rekognition.
|
65 |
+
|
66 |
+
## 🏆 Rendimiento del Modelo
|
67 |
+
|
68 |
+
| Métrica | Valor |
|
69 |
+
|-----------------|-------|
|
70 |
+
| **mAP@50** | 0.98 |
|
71 |
+
| **mAP@50-95** | 0.91 |
|
72 |
+
| **Precisión** | 0.96 |
|
73 |
+
| **Recall** | 0.98 |
|
74 |
+
|
75 |
+
### Curvas de Rendimiento:
|
76 |
+
|
77 |
+
**Curva F1-Confianza:**
|
78 |
+

|
79 |
+
|
80 |
+
**Curva Recall-Confianza:**
|
81 |
+

|
82 |
+
|
83 |
+
**Curva Precisión-Confianza:**
|
84 |
+

|
85 |
+
|
86 |
+
**Matriz de Confusión:**
|
87 |
+

|
88 |
+
|
89 |
+
## 📸 Ejemplos de Detección
|
90 |
+
|
91 |
+
Imágenes detectadas correctamente:
|
92 |
+

|
93 |
+

|
94 |
+
|
95 |
+
Ejemplo de detección errónea (mujer rubia identificada incorrectamente como Jennifer Aniston):
|
96 |
+

|
97 |
+
|
98 |
+
---
|
99 |
+
|
100 |
+
¡Déjanos saber tus comentarios o mejoras! 🚀
|
101 |
+
|
images/.gitattributes
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
*.png filter=lfs diff=lfs merge=lfs -text
|
images/F1_curve.png
ADDED
![]() |
Git LFS Details
|
images/P_curve.png
ADDED
![]() |
Git LFS Details
|
images/R_curve.png
ADDED
![]() |
Git LFS Details
|
images/confusion_matrix_normalized.png
ADDED
![]() |
Git LFS Details
|
images/image1.png
ADDED
![]() |
Git LFS Details
|
images/image2.png
ADDED
![]() |
Git LFS Details
|
images/image3.png
ADDED
![]() |
Git LFS Details
|
images/results.png
ADDED
![]() |
Git LFS Details
|
models/.gitattributes
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
models/best.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:63a048a7db16da856bdc1c51488ab82677859bb8412312040342fc9e7c631af3
|
3 |
+
size 6232803
|
models/last.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:246234d6aed795fb068e135bb94b7dc1fd39c08f5af4475407a23c93679984ec
|
3 |
+
size 6232803
|