---
license: apache-2.0
base_model:
- Ultralytics/YOLO26
tags:
- icicle
- roof
- facade
- building
metrics:
- precision = 0.6
- recall = 0.52
- mAP50 = 0.52
- mAP50-95 = 0.22
---

# 🧊 Icicle Detector – YOLO26x model

This model is trained for automatic detection of **icicles** on images of building roofs and facades.  
It detects **one class** – `icicle` – and outputs bounding boxes around each detected icicle.

**Architecture:** YOLO26x (custom modification based on YOLO).  
**Weights file:** `model.pt`

## Inference example

```python
from ultralytics import YOLO
import cv2

# Load the model from Hugging Face
model = YOLO("IgorKir16/icicle-detector/model.pt")

# Run detection on an image
results = model("path/to/your_image.jpg", conf=0.25)

# Visualize results
for r in results:
    im_array = r.plot()
    cv2.imshow("Result", im_array)
    cv2.waitKey(0)

# Print bounding boxes and confidence
for r in results:
    for box in r.boxes:
        x1, y1, x2, y2 = box.xyxy[0].tolist()
        conf = box.conf[0].item()
        cls = int(box.cls[0].item())
        print(f"icicle: ({int(x1)}, {int(y1)}) – ({int(x2)}, {int(y2)}), confidence: {conf:.2f}")