Skip to content

Comment on LLMD: A Large Language Model for Interpreting Longitudinal Medical Recordsparent

Comments

No other models that are public worth comparing to... Hippocratic advertises good benchmarks but that might be marketing fluff.

Have you checked out dataset building with nemotron? The nemotron synthetic data builder is quite powerful.

Moreso, check out model merging. It's possible if you merge some of your model against llama3.1 base it may perform much better.

Check out max labonne's work on hugging face

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.