using-docker.md

↤ BACK

Using Docker

Docker ensures everything is preconfigured, set up, and ready to go.

Requirements:

• Docker Desktop

New Container

Run the following terminal command from the root of this project:

docker build --no-cache -t image-processor .

Run Interactive Web App

This will start an interactive web application that runs on your local computer at http://127.0.0.1:8501

docker run --name fbc --rm -p 8501:8501 image-processor streamlit run app.py

Run CLI for Single Image Processing

You can process a single image using the cli.py script.

docker run --name fbc --rm -v "$(pwd)"/data:/fbc/data image-processor python3 cli.py --help
docker run --name fbc --rm -v "$(pwd)"/data:/fbc/data image-processor python3 cli.py /fbc/data/input/image.jpg --dest=/fbc/data/output/ --json --images

Run CLI for Batch Image Processing

To batch process everything in the data/input folder, use the batch.py script.

docker run --name fbc --rm -v "$(pwd)"/data:/fbc/data image-processor python3 batch.py --help
docker run --name fbc --rm -v "$(pwd)"/data:/fbc/data image-processor python3 batch.py --json --images

Run Docker in CUDA Mode

If you are on a high-end computer with an Nvidia Graphics Card that supports CUDA, you can add the --gpus=all to docker run to enable CUDA support. Just replace docker run with docker run --gpus=all