Stretch iPhone to Its Limit: 2GiB Stable Diffusion Model Runs Locally on Device

Re:Stretch iPhone to Its Limit: 2GiB Stable Diffusion Model ... by lll-o-lll on Wednesday November 09, 2022 @10:58PM (#33542016)

Any tips on how you “ If the face is cropped, now I know how to use the inpainting model to fill it in. If the inpainting model doesn’t do its job, you can always use a paint brush to paint it over and do an image-to-image generation again focused in that area.” using the app?

[ Parent | on HN ]

Re:Stretch iPhone to Its Limit: 2GiB Stable Diffusion Model ... by liuliu on Wednesday November 09, 2022 @11:34PM (#33542241)

I have a Twitter thread on this: https://twitter.com/liuliu/status/1587978815208407041?s=46&t...

(Note that at that time, there is an implemention bug in inpainting model that caused the weirdness that I need to manually fix)

[ Parent | on HN ]

Re:Stretch iPhone to Its Limit: 2GiB Stable Diffusion Model ... by ragazzina on Thursday November 10, 2022 @09:39AM (#33547088)

liuliu, this is simply incredible.

Were you focused on just making it work on the iPhone, or do you think you will keep adding functionalities to the app? Do you think it will ever be possible to train one's own model on an iPhone?

[ Parent | on HN ]

Re:Stretch iPhone to Its Limit: 2GiB Stable Diffusion Model ... by liuliu on Thursday November 10, 2022 @10:35AM (#33547944)

I think that fine-tuning the whole model (a.k.a. Dreambooth) on iPhone would require more RAM / processing power than it currently has. More viable path is to implement Hypernetwork + Textual Inversion, that is within possibility of today's hardware.

[ Parent | on HN ]