This repository contains the code for a large language model (LLM) that was built for radiologists. It can be used to answer questions about skeletal injuries , generate radiology reports, and more.
- Langchain
- Huggingface
- FAISS
- OpenAI
- NLTK
Before running the notebook , upload the Osborn Brain Imaging textbook to the drive or download it . Textbook drive link : https://drive.google.com/file/d/1h-6wvlOx9vVuHimUmWOrHNk36lm4lOjd/view?usp=drive_link
If the above mentioned textbook added to the drive , change pdf_paths to :
pdf_paths = ["/content/drive/MyDrive/Osborn Brain Imaging.pdf"]If downloaded locally change the path accordingly.
- QA with Vectorised Textbook information
- Remembers context of conversation
- Few Shot Answering
To run this project, you will need to add the following environment variables to your .env file
OPENAI_API_KEY :
HUGGINGFACE_API_KEY :
Now run the notebook!