Clone this repo to spin up an ML web app using React, Flask and Torchserve.
Edit the .env file to include your AWS credentials, the ${AWS_MODEL_STORE} S3 bucket and the ${MODEL_NAME}.
The Dockerfile downloads the state dict from s3://${AWS_MODEL_STORE}/${MODEL_NAME}_state.pt. Read more about configuring torchserve at https://github.com/pytorch/serve
Spins up an API server that receives inputs from the frontend, and returns predictions from the model server.
Serves an nginx proxy that interfaces with the flask API server. JS files are in frontend/src and was populated using create-react-app.
run docker-compose up