I noticed that your paper mentions that this model can be applied to video editing, is there any code for this?