2 ms·Learned Similarity Autoencoder for Modelling and Reconstructing Video Frames1 points by joubert 10y ago