The goal of this project is to create a question answering demo using OpenAI embeddings and a sample text.
For more information, please read the project roadmap and the solution proposal.
- Unix (Ubuntu, WSL 2, etc.)
- Python > 3
- Pip >= 22
- Venv > 3
You should already have these dependencies installed.
Install the python dependencies:
bash manage.sh install
Generate synthetic questions and embeddings:
bash manage.sh prepare "io/data/sample.txt"
Note: You must pass the path to a text file as the first argument.
Train a model to answer questions:
bash manage.sh train
Ask a question through the command line:
bash manage.sh ask "What is the name of the main character?"
Note: Please use double quotes to pass the question.
Start a demo webserver:
bash manage.sh serve