Dockerize-Model is an end-to-end, Docker-ready deep learning project that classifies fresh vs. rotten fruits (peaches, pomegranates, strawberries). The repository includes data processing utilities, model training code, a FastAPI inference service, and Docker configuration for reproducible deployment.

• Features
• Quick Start
• Local Development
• Docker Deployment
• API Reference (examples)
• Configuration
• Project Structure
• Notes: Model & Data
• Contributing
• License

• Image classification for multiple fruit types (fresh/rotten)
• Modular and well-organized code (data, models, api, utils)
• FastAPI REST service for inference
• Dockerfile for consistent deployment
• Logging and configuration via src/constants/constants.py

1. Build the Docker image:

docker build -t dockerize-model .

1. Run the container and expose the API:

docker run -it -p 5000:5000 dockerize-model

1. Visit the interactive API docs: http://localhost:5000/docs

1. Create and activate a virtual environment (recommended):

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

1. Run the FastAPI app locally:

python src/api/main.py

1. Open the API docs: http://localhost:5000/docs

Predict endpoint:

• POST /predict
  • Accepts: multipart/form-data with a file field named file
  • Returns: JSON with FileName, PredictionClass, and prediction (confidence)

Example curl request:

curl -X POST "http://localhost:5000/predict" -F "file=@/path/to/image.jpg"

Example Python client (requests):

import requests

url = "http://localhost:5000/predict"
with open("/path/to/image.jpg", "rb") as f:
    r = requests.post(url, files={"file": f})
print(r.json())

Sample response:

{
  "FileName": "strawberry.jpg",
  "PredictionClass": "Fresh Strawberries",
  "prediction": 0.9876
}

• Most runtime/configuration values live in src/constants/constants.py (e.g., MODEL_PATH, HOST, PORT, and RESIZE).
• Make sure the model file is available at models/checkpoints/model.keras or update MODEL_PATH.

Note: The README previously referenced src/constant/constant.py. The correct path is src/constants/constants.py.

Dockerize-Model
├── Dockerfile
├── README.md
├── requirements.txt
├── src/
│   ├── api/           # FastAPI app, endpoints and schemas
│   ├── constants/     # Paths and configuration
│   ├── data/          # Data ingestion/processing helpers
│   ├── models/        # Model definition / training utilities
│   └── utils/         # Helpers (e.g., model loading)
├── data/
│   ├── raw/
│   └── processed/
├── models/
│   └── checkpoints/   # Place `model.keras` here
├── notebooks/         # EDA and experiments
└── logs/

• Pretrained model: models/checkpoints/model.keras (expected by default). If not present, inference will fail when loading the model — see logs for details.
• Data files used for experiments are in data/processed/*.json (train/test splits).

• If server fails to start, check MODEL_PATH and ensure the model file exists.
• Validate image uploads via the /docs UI to see expected request shape.
• Logs are written to the logs/ directory (see LOG_FILE and LOG_DIR in src/constants/constants.py).

Contributions, bug reports, and improvements are welcome. Please open an issue or submit a pull request.

This project is licensed under the MIT License — see the LICENSE file.