Questions about the differences between two training scripts.

jy0205 / Pyramid-Flow

Code of Pyramidal Flow Matching for Efficient Video Generative Modeling

https://pyramid-flow.github.io/

MIT License

2.4k stars 233 forks source link

Questions about the differences between two training scripts. #164

Closed masterwu2115 closed 5 days ago

masterwu2115 commented 2 weeks ago

Hi authors, thanks for the well-written code, cause I'm a novice in video generation, I'm curious about the difference between the two scripts you provided, i.e., scripts/train_pyramid_flow.sh and scripts/train_pyramid_flow_without_ar.sh. It seems like the first one is for t2v and the second one is for t2i? Does that mean I need to fine-tune separately for t2i and t2v?

feifeiobama commented 2 weeks ago

Both are for t2v. The first is our proposed method (spatial pyramid + temporal pyramid), the second is just the spatial pyramid.