rotten-work/simple-cat-detector

A simple cat detector using neural networks and deep learning based on VGG using Tensorflow Serving with Docker

Simple Cat Detector

A simple cat detector using neural networks and deep learning based on VGG using Tensorflow Serving with Docker.

Documentation

Set up

• Install Anaconda
• Install Tensorflow
• Enable the WSL 2 feature on Windows
• Install Docker Desktop on Windows
• Install Tensorflow Serving using Docker

Use

Save the VGG model to disk:

python save_vgg_model.py

Start TensorFlow Serving container and open the REST API port using a WSL terminal:

# Change the '/mnt/d/GitHub/' to your own repository path
docker run --name vgg_test -t -p 8501:8501 --mount type=bind,source=/mnt/d,target=/mnt/d -e MODEL_NAME=vgg_serving -e MODEL_BASE_PATH=/mnt/d/GitHub/simple-cat-detector tensorflow/serving &

Query the model using the predict API in an Anaconda Powershell Prompt:

python request_vgg_serving.py