Open
Description
Hi, @coryMosaicML
Thank you for merging the recent PR #155 related to Diffusion Transformer (DiT). I have a couple of questions regarding the use of DiT in this repository:
- Is there a specific configuration available for running DiT within this repository?
- Similar to SD2 and SDXL, is it necessary to train DiT across multiple resolutions in a staged manner?
Your guidance on these points would be greatly appreciated.
Metadata
Metadata
Assignees
Labels
No labels