greenelab / tybalt

Training and evaluating a variational autoencoder for pan-cancer gene expression data
BSD 3-Clause "New" or "Revised" License
164 stars 62 forks source link

write scripts/vae_pancancer.py #17

Closed gwaybio closed 7 years ago

gwaybio commented 7 years ago

this will be the workhorse of the repository - will need to take in command arguments and run through model training and preliminary evaluation.

By preliminary evaluation, I mean it will need to determine quickly which parameter combination is best, and output some sort of metric. I am thinking the metrics should be a combination of lowest loss and biological interpretability.