About appling the strategy to DiT

[Jianxff]

Thanks for your great work.

I'm trying to use the structure control strategy in DiT-based diffusion model.
However, DiT blocks are different from UNet encoder-decoder, do you have any suggestions about applying to DiT?

Thanks again.

[SichengMo]

Thanks for reaching out! We have tested this method on transformer-based diffusion models, and simply following our approach can lead to great generation results. However, we have empirically found that this method does not work well on Diffusion Transformers with an AdaIN module. You may want to start with a diffusion model without an AdaIN module, such as Bagel.