Senu-12/snapstock-fruit-vegetable-detector

SnapStock-AI Fruit and Vegetable Detector

This repository stores a YOLOv8m fruit and vegetable object-detection checkpoint used as a baseline model for the SnapStock-AI university project.

Original source

The checkpoint was obtained from the baseline models published with:

• Original author: Henning Heyen
• Original project: Fruits-And-Vegetables-Detection-Dataset
• Dataset: LVIS Fruits and Vegetables
• Classes: 63
• Architecture: YOLOv8m

This repository does not claim authorship of the original dataset or baseline model. The checkpoint is mirrored here to provide stable model downloading for the SnapStock-AI development team.

Intended use

The model is used to:

• Detect fruits and vegetables
• Return bounding boxes
• Identify object classes
• Count detected items

Limitations

The original project reports class imbalance and duplicate Tomato/tomato and Strawberry/strawberry categories. Performance should be tested using real SnapStock-AI inventory images before production use.

Local loading

from ultralytics import YOLO

model = YOLO("fruit_vegetable_yolov8m.pt")
results = model.predict("inventory-image.jpg")