search

ldeo-glaciology / LEAP-Cryo-planning

A repo for planning and tracking progress on the LEAP-Cryo project: Learning ice-sheet flow with physics-based and machine learning models.
2 stars 2 forks source link

VAE demo #20

Closed Templar129 closed 1 year ago

Templar129 commented 1 year ago

Try to use TG data as input then train a Variational Autoencoder.

Templar129 commented 1 year ago

I now have a really early version of VAE. It is working now and can use a batch of about 20-50 high resolution image as training data. Right now I am using duplicate of the original data as training data, so that means I use 20 identical image as input. It now can produce a variety of images. Once I find out a suitable way to generate training data, I believe this model could be helpful.

Below is a few demos created by the model: image image image