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