This model predicts the next frame for a time series of images. A sample of 10 frames associated with the moving MNIST dataset is shown below. The original sample at one time step vs. predicted is shown below. Original: Predicted: