Adaptation to FLUX models?

Open Shyryp opened this issue 11 months ago • 1 comments

Is there a way to adapt this technology to FLUX 1.dev models?

Currently it is not possible to use these nodes for FLUX due to the size difference VAE.

Thanks for info!

Jan 26 '25 00:01 Shyryp

just wondered as well if anyone has trained a flux version of this. my understanding is this repo here has the same purpose as https://github.com/Ttl/ComfyUi_NNLatentUpscale

Mar 08 '25 14:03 tazztone