This repo shows a minimal example to build your web app with Streamlit + Tensorflow. It has been designed following the microservices architecture. Two services are available:
- The front end of the app is build with Streamlit
- The back end deploys a pretrained Tensorflow model as servable thanks to Tensorflow Serving.
Once the app is running you can pass the link of an image from the web. Tap on the Classify button and it tells you immediately the class of the image 💫
It is assumed you have Docker installed. Then:
- Clone this repository with
git clone https://github.com/mlgxmez/streamlit_tf_app.git - Go to the root folder of the repo and type in your terminal
docker-compose up -d - Open the web browser and visit
localhost:8503
The web app is now running in your computer. It's up to you how you want to extend the project. I would encourage you to read the Dockerfiles and the script running the app to see how all the files work together ⚙️
- To shut down your containers at once type
docker-compose down