Open SkywalkerSZW opened 2 years ago
I use moving-gif-128 dataset and moving-gif.yaml you provide to train a model on my GPU, but my model performs much worse on video generation than the model you provided. What training strategy did you adopt to make the model work so well?
I guess you better try this repository https://github.com/snap-research/articulated-animation
I use moving-gif-128 dataset and moving-gif.yaml you provide to train a model on my GPU, but my model performs much worse on video generation than the model you provided. What training strategy did you adopt to make the model work so well?