Best local setup to summarize ~500 pages of OCR’d medical PDFs?
I have about 20 OCR’d PDFs (~500 pages total) of medical records (clinical notes, test results). The OCR is decent but a bit noisy (done with ocrmypdf on my laptop). I’d like to generate a structured summary of the whole set to give specialists a quick overview of all the previous hospitals and exams.
The machine I can borrow is a Ryzen 5 5600X with an RX 590 (8GB) and 16GB RAM on Windows 11. I’d prefer to keep everything local for privacy, and slower processing is fine.
What would be the best approach and models for this kind of task on this hardware? Something easy to spin up and easy to clean up (as I will use another person's computer) would be great. I’m not very experienced with local LLMs and I don’t really feel like diving deep into them right now, even though I’m fairly tech-savvy. So I’m looking for a simple, no-frills solution.
TIA.