BARS.AI : Colon Cancer Chat Bot

Inspiration

Samrudh's family friend had colon cancer at the age of 50 and had a doctor unable to diagnosis him.

What it does

Our model takes the dataset of colon cancer and breaks the information into chunks to be analyzed and provides a diagnoses and symptoms

How we built it

The code is designed to extract text from PDF document, process it into manageable chunks, generate embeddings for these chunks, and then use these embeddings to answer user queries based on the most relevant passage from the document. The Flask application provides endpoints to interact with functionality.

Challenges we ran into

We had a problem working with our UI. Our web application was not imaging. We also had problems with our python code as it was not able to store the amount of tokens we had. We had to shorten our training model count.

Accomplishments that we're proud of

We are proud that we were able to build an LLM that is able to be a benefit to the medical community and those at home.

What we learned

We learned that creating a LLM is lengthy and challenging but at the end of the day it is worth it because it is beyond anything we initially expected that we could do.