18 projects
MEDS-EIC-AR
A simple auto-regressive, 'everything-is-code' style model for MEDS datasets
MEDS-transforms
A framework for compiling simple, mapreduce style pipelines over MEDS datasets.
dftly
dftly (pronounced deftly) is a simple library for a safe, expressive, config-file friendly, and readable DSL for encoding simple dataframe operations.
MIMIC-IV-MEDS
An ETL pipeline to extract MIMIC-IV data into the MEDS format.
yaml-to-disk
A simple utility to define directory structures in YAML and populate them on disk in a single command.
MEDS-extract
MEDS ETL building support leveraging MEDS-Transforms.
MEDS-trajectory-evaluation
Utilities for evaluating autoregressive generated trajectories over MEDS datasets, including temporal AUC computation and zero-shot ACES task labeling.
meds-torch-data
An efficient, flexible PyTorch dataset class for MEDS data.
nested-ragged-tensors
Utilities for efficiently working with, saving, and loading, collections of connected nested ragged tensors in PyTorch
ml-mixins
A collection of useful mixins for machine learning development code.
MEDS-visualizations
A framework for compiling simple, mapreduce style pipelines over MEDS datasets.
meds-testing-helpers
Builds sample MEDS datasets for testing.
flexible-schema
A simple class to aid in defining flexible schemas for PyArrow datasets.
eICU-MEDS
An ETL pipeline to extract the eICU dataset into the MEDS format.
pretty-print-directory
A simple utility to pretty-print a directory tree, suitable for use in pytest test cases.
MEDS-DEV
Task configuration and helper files for the MEDS-DEV Benchmark
hydra-profiler
A simple hydra profiler to track and record memory usage and runtime information of jobs.
pytorch-lognormal-mixture
A pip installable version of the lognormal mixture distribution from https://github.com/shchur/ifl-tpp/tree/master/code
