Denoising strength

[alelordelo]

Thanks for sharing this great repo!

Is it possible to control Denoising strength? I know by default the app doesn’t support, but does the coreML SD package support it?

[pcuenca]

Hi @alelordelo! What do you mean exactly by denoising strength? I've seen that term in the context of image-to-image generation, but so far we are doing text-to-image. Would you mind clarifying a bit? :)

[alelordelo]

Hi Pedro,Yup, that is a parameter that affects heavily image to image generation.So if there is only text to image now, no need to add it.Thanks!Sent from my iPhoneOn 1 Mar 2023, at 20:12, Pedro Cuenca ***@***.***> wrote: Hi @alelordelo! What do you mean exactly by denoising strength? I've seen that term in the context of image-to-image generation, but so far we are doing text-to-image. Would you mind clarifying a bit? :) —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: ***@***.***>

[pcuenca]

Cool, understood! This repo does not have image to image (yet), but it was recently added to Apple's one (which is a dependency here). See here for the strength parameter: https://github.com/apple/ml-stable-diffusion/blob/main/swift/StableDiffusion/pipeline/StableDiffusionPipeline.Configuration.swift#L27

[alelordelo]

Ah, nice! Thanks for pointing that!Sent from my iPhoneOn 2 Mar 2023, at 10:18, Pedro Cuenca ***@***.***> wrote: Cool, understood! This repo does not have image to image (yet), but it was recently added to Apple's one (which is a dependency here). See here for the strength parameter: https://github.com/apple/ml-stable-diffusion/blob/main/swift/StableDiffusion/pipeline/StableDiffusionPipeline.Configuration.swift#L27 —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: ***@***.***>