ehrmonize is package to abstract medical concepts using large language models.
Project description
EHRmonize
Welcome to ehrmonize
, a Python package to abstract medical concepts using large language models.
Citation
TBD
As a data scientist working with EHR data, I want to be able to automatically categorize medical concepts, so that I can collaborate more efficiently with clinicians.
Rationale
Corpus generation from EHRs provides the
data that needs categorization, across different domains and tasks, which is then fed to ehrmonize
to employs LLMs to categorize the entries into predefined classes.
Current supported tasks
Type | Task |
---|---|
Free-text | get_generic_name |
get_generic_route | |
Multiclass | classify_drug |
Binary | one_hot_drug_classification |
Custom | custom |
Current supported models / engines / APIs
API | model_id |
---|---|
OpenAI | gpt-3.5-turbo |
gpt-4 | |
gpt-4o | |
AWS Bedrock | anthropic.claude-instant-v1 |
meta.llama3-70b-instruct-v1:0 | |
mistral.mixtral-8x7b-instruct-v0:1 |
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
ehrmonize-0.1.0a1.tar.gz
(6.5 MB
view hashes)
Built Distribution
Close
Hashes for ehrmonize-0.1.0a1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | fdd2a88e447bfd9e3b450189d0b76a6ed2ef96939646d0c460612fb8dc316569 |
|
MD5 | 095c44b6b7d9aee84fe40794c3e4ae27 |
|
BLAKE2b-256 | d87e3a82ca1c52f4ee086f8f5d9c2ccf7dd2087905ab4e1b797fb25f8bdd0da7 |