Skip to content

Medplexity

Release Documentation Status Discord License: MIT Open in Colab

Medplexity is a python library to help with evaluation of LLMs for medical applications.

It provides easy interfaces to use existing benchmarks, including MedQA, PubMedQA, etc. and evaluate performance of LLMs and more complex chains on these benchmarks.

Who should use Medplexity?

We see medplexity most useful for the following groups:

  • Builders in the medical space who want to experiment with different architectures.
  • Researchers who want to evaluate their new LLMs on existing benchmarks.
  • Medical professionals who want to understand the capabilities of LLMs and their limitations.
  • Anyone curious about LLMs in healthcare :)

Installation

To get started with medplexity, you first need to install it. You can do this by running the following command:

# Using poetry (recommended)
poetry add medplexity

# Using pip
pip install medplexity

Examples

See an example use with OpenAI and MedQA dataset in MedQA notebook.

Community & Support

Please join our Discord for any questions, support, and just to share what you are working on and what could be beneficial for you.

License

Medplexity is licensed under the MIT License. See the LICENSE file for more details.