HealthCareLLM- An LLM which makes sure patients are guided swiftly and efficiently through responses generated by a collaboration of information from:
-
Their diet chart meals(to ensure they are eating the right things in specific meals.
-
The general notes assigned to them by the doctor(specific to each patient).
-
The notes assigned with each meal(to ensure that the meal is taken the correct way).
-
The timing associated with each meal(so that they have their meals at the right time).
Model used- Llama 3.1 from Ollama(locally). Even more accuracy and closeness to the ideal_responses can be achieved by using paid APIs. This project is solely based on use of Open Source LLMs.
I have run 2 different codes- just to ensure the response is better. Final code is in new.ipynb file and the final output is in final_output.json file.
STEPS TO RUN:
- Just create a virtual environment in VS Code and run the ipynb file. Simple.
BONUS: I tried to make it act like your family doctor! See responses in family_doctor.json file.