https://huggingface.co/AbstractPhil/ My encoder sampler project is here, very powerful stuff. https://github.com/AbstractEyes/comfy-clip-shunts/tree/dev2_electric_boogaloo I'll be taking a break from large scale training and instead shift my focus to releasing regular small train loras to augment the currently existing models that I have; after the current SD35 training concludes. I'll also be spending more time refining my tools like the clip suite and the unreleased caption software I always babble about. I must prepare... for the big train. I have fiber, a lot of data, and I'm working on a flexible Flux based model meant to be a cousin to Illustrious, Noob, Pony, or any of the other NSFW/SFW models. Flux is a costly beast to train so the process is slower than training a similar SDXL based model.