search

ermongroup / ddim

Denoising Diffusion Implicit Models
MIT License
1.48k stars 211 forks source link

Transferability to transformers #16

Closed anthony-mendil closed 1 year ago

anthony-mendil commented 2 years ago

First of all, thank you for providing the code! In the according paper I read that the only difference between DDPM and DDIM is how samples are generated. Intuitively, I would then assume that the CNN model could also be replaced by a transformer-type architecture. If my understanding of your paper is correct, one could still use the same simple loss and your DDIM sample generation. I would highly appreciate your opinion on this. Thanks in advance, Anthony.

jiamings commented 1 year ago

Yes. Although I have no idea how a transformer architecture would work on this,

anthony-mendil commented 1 year ago

Thanks for you answer. Could you clarify what you mean by that statement? Are you referring to how well it works or do you have concerns regarding the technical implementation?