Is it possible to use VersatileDiffusion for Seq2Seq generation training?

[HZQ950419]

Hi,

Thanks for the amazing project for diffusion models! I'm interested in Seq2Seq generation with diffusion models. Thus is it possible to use VersatileDiffusion model for Seq2Seq generation training? Or are you planning to add text-to-text generation pipeline in Diffusers?

Thanks!