Dear authors,
After reading through your paper, I have some questions about the training dataset and method.
Could you please provide the sources of the original dataset for training FramePack? Also, did you do full-parameter training of the base Hunyuan/Wan model to get the framepack model or just train certain layers? Thanks a lot for your help!