For deploying the model to an AWS service, we have to create API Endpoints for routing. Flask is being used for that.
We are currently facing issues for developing a Flask server locally, and on Google Colab and Kaggle due to dependency depreciations.
Thus, currently, the model is integrated to the client's user interface using ngrok. Ngrok is a globally distributed reverse proxy for securely accessing applications being processed on IP Addresses of a cloud service to our local IP Addresses. It creates secure tunnels from a public internet endpoint to a locally running server. This endpoint is then accessed using the client interface built using Streamlit where the model is integrated to allow users for interacting with our machine learning model.
Flask
is being used for that.ngrok
. Ngrok is a globally distributed reverse proxy for securely accessing applications being processed on IP Addresses of a cloud service to our local IP Addresses. It creates secure tunnels from a public internet endpoint to a locally running server. This endpoint is then accessed using the client interface built usingStreamlit
where the model is integrated to allow users for interacting with our machine learning model.