Text to DyNCA

This code extends the work done in [DyNCA] to create NCAs able to synthesize animated pictures from text. The repo contains also a motion prediction neural network that predicts the vector flow target of the DyNCA.

To train the model, please use colab by clicking on the button below

Run in Google Colab

Train a motion prediction UNet

You can achieve this by runing the notebook [here]

Florian Constantin, Julien Sahli, Jad Sobhie