I am working to develop a Q&A chatbot model which can answer all medical related queries. I have limited computing resources. Can someone guide me on how to build the model using any pretrained models which are already trained on huge medical database since I don't have much computing resource to train the model on huge datasets?
#Help with developing medical Q&A model
celest ibex
spark pewter
I just finished helping someone get resources for a retrieval based Q&A task with documents and a pretrained LLM model. You can try and see if you can adapt the example to your use case as a start.
Fair warning, this will probably work best on a linux env because it uses the faiss package in python but feel free to read more documentation on the vector stores. I hope your resources are at around a consumer grade gaming PC with decent GPU because this will take most of your resources. if you have trouble running these models on your device, consider investing in a cloud instance until you have funds for a more powerful machine or the project is over.
The idea here is that the model uses the documents which are embedded into a vector store database to influence its answers, so you dont need to retrain the model, only update your database.