LANTERN webUI

Streamlit UI for training NIST LANTERN machine learning model.

Prerequisites

1. Install LANTERN on the deployment system (see Instructions from LANTERN repo).
2. Install streamlit on the deployment system:

   pip install streamlit
   

Deployment and Usage

streamlit run lantern_server.py

will launch a webUI running at http://localhost:8501.

Usage

1. Drag and drop CSV phenotype data with column labels = phenotypes, variance ('phenotype_var').
2. Select desired phenotypes for training from dropdown menu. Matching variance columns are automatically selected.
3. Choose the number of epochs to train for (standard 1000 in publications).
4. Specify an output directory on the server (ie. output/myrun1)
5. Click 'Start Training'
6. When complete, download model files (model.pt, loss.pt) using UI or copy from output directory.