Deploy BERT for Sentiment Analsysis with FastAPI

Deploy a pre-trained BERT model for Sentiment Analysis as a REST API using FastAPI

Local Usage

Clone this repo:

git clone https://github.com/kandeldeepak46/Fine-Tuning-BERT-For-Sentiment-Analysis-Served-With-FastAPI.git
cd Fine-Tuning-BERT-For-Sentiment-Analysis-Served-With-FastAPI

Install the dependencies:

pipenv install --dev

Download the pre-trained model:

bin/download_model

Test the setup

Start the HTTP server:

bin/start_server

if everything goes well, the fastAPI sucessfully get started

Demo

The model is trained to classify sentiment (negative, neutral, and positive) on a custom dataset from app reviews on Google Play. Here's a sample request to the API:

 curl -d "{\"text\":\"This game is amazing, it is literally part of my childhood. It works well with hand eye coordination, and might even help with reflexes (not positive, just a guess)This game can keep you interested for hours,and has a lot of small things to work for! I really like the way the game has been moving as of update.\"}" -X POST http://localhost:8000/predict

The response you'll get looks something like this:

{
    "probabilities": {
        "negative": 2.0558945834636688e-05,
        "neutral": 4.625277506420389e-05,
        "positive": 0.9999332427978516
    },
    "sentiment": "positive",
    "confidence": 0.9999332427978516
}

OR

Send a test request via shell:

bin/test_request

!!!V O I L A!! There we go

License

MIT