model-inference-api

Serving the Machine Learning model through API for inference.

Warning

Authorization functionality is not yet implemented.

DO NOT USE IN PRODUCTION!!!

Necessary Logic

Prediction Logic

• Record input text from the client
• Pass the text through preprocessing.py
• Tokenize the text for BERT-based model
• Make prediction
• Send the predicted label back to the client

Feedback Logic

• Take user feedback on the predicted label
• Store the text, label along with user feedback in the database

Endpoints

Endpoint	Status	Objetive
/predict	Implemented	Takes text as input through POST request and spits out the predicted label for the text.
/feedback	Implemented	Records user feedback on the predicted labels through POST request and stores it in database.

Security

• Rate Limits implemented for every endpoint.