Deploying a Model with TensorFlow Serving and Flask

This project demonstrates how to deploy a pre-trained TensorFlow model using TensorFlow Serving (with Docker) and create a Flask-based web interface for making predictions.

Steps to Run

1. Start TensorFlow Serving with Docker

• Open a terminal or Command Prompt (as administrator on Windows).
• Run the following command:

docker run -p 8501:8501 --name=pets -v "C:\pets:/models/pets/1" -e MODEL_NAME=pets tensorflow/serving

Note: Replace C:\pets with your model directory path.

2. Set Up and Run Flask

• Create a virtual environment:

python -m venv flaskapp

• Activate the virtual environment:

source flaskapp/bin/activate

• Install dependencies:

pip install -r requirements.txt

• Start the Flask app:

python app.py

3. Open the Web App

Go to http://localhost:5000 in your browser to use the app.

Requirements

Install all Python dependencies with:

pip install -r requirements.txt