Published December 19, 2022 | Version v1
Journal article Open

Trained neural networks for (probabilistic) ECG regression of electrolytes

Creators

  • 1. Tübingen University
  • 2. Uppsala University
  • 3. Uppsala Univserity
  • 4. Uppsala University, University of New South Wales

Description

Pre-trained pytorch models accompanying the source code at https://github.com/philippvb/ecg-electrolyte-regression.

 

Available models are:

  • models for regression of all electrolytes. The model weights used for the graphics in the paper are the ones with the median performance. That is model 3 for potassium; model 4 for calcium; model 2 for creatinine; and model 4 for sodium.
  • model for probabilistic regression of all electrolytes.

Files

models.zip

Files (7.5 GB)

Name Size Download all
md5:edf678ceeababb5f739c1b09d2ce96ac
7.5 GB Preview Download