Open TaoshuaiZ opened 8 months ago
Thank you for your interest in our benchmark! I have just released implementations and pre-trained weights for the FitVid, SVG', MCVD, and Struct-VRNN models, and updated the repository README with instructions on how to download and use them. I have also reorganized some of the configurations to make it easier to directly download the files and start running the sample code. Please let me know if you have any further questions or concerns!
I've noticed that the reimplementation codes for video prediction models (i.e., FitVid, SVG', MCVD, MaskViT, Struct-VRNN) are missing. Do you plan to release these codes and the pre-trained weights for them? I believe providing these reimplementation codes would greatly enhance the utility of your benchmark.