Inspiration
We've always had a passion of using technology to help out the healthcare industry. Creating this AI Nutrition Chatbot was a way to use our passion for healthcare and technology for the good of the community.
What it does
It takes in an ingredient, vitamins that the user wants, minerals that the user wants, protein, fats, and carbs and it generates meals that the user can have depending on their dietary preferences.
How we built it
I first cleaned a dataset that contained a list of ingredients and their health quantities using jupyter notebook, and then we used app.py to create a Flask app so that the user can interact witht the computer. We then integrated GPT-2 to make the responses AI-generated, and we used frontend tools to make our chabot user-friendly.
Challenges we ran into
Some challenges that we ran into was figuring out which GPT model to use because not all of them are open source. We also were struggling with the frontend, as only one of us knew the frontend. However, one of us was able to learn frontend a little bit that they were able to debug the code that the person who knew frontend wrote.
Accomplishments that we're proud of
We're proud of having a working chatbot and that we were able to use SQLite to create a login page and our usage of GPT-2 makes our chatbot highly technological.
What we learned
We learned how to connect frontend and backend technologies using Flask and make a user-friendly interface and also using GPT as a technology in our chatbot.
What's next for AI Nutrition Chatbot
We will try to make our responses more optimized by looking for a better GPT model and enchancing our SQLite database.
Log in or sign up for Devpost to join the conversation.