Update README.md

README.md

@@ -1,39 +1,30 @@
 ---
-title: AegisRoad Inference
-emoji: 🛣️
-colorFrom: red
-colorTo: green
+title: AegisRoad — Road Damage Detection
+emoji: ⚡
+colorFrom: blue
+colorTo: red
 sdk: docker
-pinned: false
 app_port: 7860
+pinned: true
+short_description: YOLOv11 road damage detection · D00/D10/D20/D40 · RDD2022
 ---
 
-# AegisRoad v3.0 — Road Damage Detection API
+# ⚡ AegisRoad v3.0
 
-YOLOv8-Nano inference server for detecting road damage in dashcam images.
+AI-powered road damage detection using YOLOv11.
+
+## Endpoints
+- `POST /predict` — upload an image, get detections + annotated image back
+- `GET /health` — check if model is loaded
+- `GET /classes` — list damage classes
 
 ## Classes
-| Code | Damage Type | Severity |
-|------|-------------|----------|
+| Code | Name | Severity |
+|------|------|----------|
 | D00 | Longitudinal Crack | Low |
 | D10 | Transverse Crack | Medium |
 | D20 | Alligator Cracking | High |
 | D40 | Pothole | Critical |
 
-## API Endpoints
-
-- `GET /` — Service info
-- `GET /health` — Health check
-- `GET /classes` — List all damage classes
-- `POST /predict` — Upload image, get detections
-
-## Usage
-
-```bash
-curl -X POST "https://YOUR-SPACE.hf.space/predict" \
-  -F "file=@road_image.jpg"
-```
-
-## Model
-Upload your trained `best.pt` from Kaggle into the `models/` folder.  
-Without the model, the API runs in **demo mode** with simulated detections.
+## Activating Real Inference
+Upload `best.pt` to the `models/` folder after Kaggle training. Space auto-restarts.