-
two components:
- world model (vision model CNN + memory RNN)
- controller model
-
vision model is a VAE, compress image to small latent vector z
You can use the generate_vae_data.py
script to generate a dataset for training the VAE model. Before you do this, make sure you have enough space on your disk (~260 GB).
After you have generated your training dataset, you can use it to train your Variational Autoencoder Model.
- add image transforms for the VAE training