MedCruncher - Automated labeling of medical reports and documents using LLMs (Large Language Models)

About

MedCruncher is a python based a wrapper to process the medical literature using Large Language Models.

Figure 1.0 MedCruncher takes multiple reports and prompt in form of excel sheet and processes the reports and prompts resepectively and saves the results into excel format

Instructions

Step 1: Install MedCruncher using pip

pip install medcruncher

Step 2: Install Fastchat for LLM inference

pip install fschat

Step 3: Follow the Fastchat's documentation Open-AI compatible api

https://github.com/lm-sys/FastChat/blob/main/docs/openai_api.md#restful-api-server

step 4: Example code to use MedCruncher module:

from wordscruncher import Cruncher
instance = Cruncher("reports.xlsx", "EMPTY", "http://localhost:8000/v1", "vicuna-7b-v1.5-16k", "final_results")
instance.crunch_data()