rohanhbtu / hinglish-fastapi

Hinglish-Transformer

It uses a Encoder-Decoder architecture with Bahdanau attention mechanism using GRU RNN cell. This has been tested and resulted with a mean BLEU score of 50%. It has been Deployed using FastAPI Framework.

To setup this repo into your local machine:

1. Fork the repository
2. Open git bash and clone it using git clone https://github.com/your_username/Hinglish-FastAPI.git
3. Create a virtual environment using py -m venv env in Command Prompt (Note: Make sure you have pip installed virtualenv package)
4. Activate the environment using .\env\Scripts\activate
5. Pip install all the python packages mentioned in the requirements.txt using py -m pip install -r requirements.txt
6. With Command Prompt opened in present working directory
7. Execute: uvicorn app:app --reload

Your app will start at a localhost port(only for development).