YOLOv8 Fruit and Vegetable Intelligent Detection System

An English-only packaging of the Streamlit-based YOLOv8 app for fruit and vegetable detection. Includes pretrained weights, custom weights, and the prepared dataset.

Quick Start

• Ensure Python 3.8+ is installed.
• From the Source_Code folder, install dependencies:

  pip install -r requirements.txt

• Launch the UI:

  python main.py

  (equivalent: streamlit run app.py)
• Optional training:

  python train.py

Layout

• app.py / main.py: Streamlit interface entrypoints.
• train.py: YOLOv8 training using dataset/fruit_veg_v1/data.yaml.
• prepare_data.py: Rebuilds the dataset from COCO archives in downloads/.
• models/: yolov8n.pt, yolov8s.pt, and custom weights (best.pt, last.pt).
• dataset/fruit_veg_v1/: images, labels, and data.yaml (English class names).
• packages.txt: system packages for OpenCV on Streamlit Cloud (libgl1, libglib2.0-0).

Notes

• Run commands from Source_Code so relative paths to models and dataset resolve correctly.
• For GPU acceleration, install the CUDA build of PyTorch matching your CUDA toolkit.
• Keep Documentation/ at the project root for your PDF/HTML manuals when packaging for Codester.