My Contribution:
Runway ML Image Training
Discript Audio Training
Jibber was created from a series of data sets we recorded in and around the Wind Tunnel These data sets include portraits (individual portraits of us) images (of the Wind Tunnel and its surroundings) text (historical documentation of the wind tunnel as well as documentation of ArtCenter) audio (recordings of our voices reading some of the text data set).
The training process of background images consists photos of MDP interior, using a bedroom preset and took 2000 steps.
The training process of the face includes three team members dataset, using a face preset and took 3000 steps