This 🫒 model is a merge of four very popular trained base SD1.5 models, it was created because of "nostalgia" for the times of Stable Diffusion 1.5, but also for some interesting experiments with ComfyUI workflows, as well as, for your delight (of course).
Models in this merge are:
✨AbsoluteReality: https://civarchive.com/models/81458/absolutereality?modelVersionId=132760
✨A-Zovya Photoreal: https://civarchive.com/models/57319/a-zovya-photoreal?modelVersionId=99805
✨LEOSAM's HelloWorld Base Model: https://civarchive.com/models/43977?modelVersionId=113623
✨MagMix: https://civarchive.com/models/18523/magmix?modelVersionId=306220
Description
some tweaks for better hands and different ethnicities
FAQ
Comments (2)
Haha, "nostalgia for the times of Stable Diffusion 1.5", I'm probably doing it wrong, but I still can't get quality out of XL like I can in 1.5. It's so resource heavy on my machine that it actually crashes Python after using it a while. LCM and Taesd vae just produce low quality even compared to fp16 1.5, and Turbo requires an entirely different setup for the same low quality. Proper anatomy and nude imagery are still wildly inconsistent, since community model developers have to train in any of that content from scratch.
1.5 isn't going away, and I have a feeling XL will go the way of 2.0 (fade into nothing) as people get tired of dealing with it. Animation and video is already stealing it's thunder, SDXL was the hot new thing for about 12 seconds, but I don't see it holding on. Could be wrong of course, but I'm personally still on team 1.5, as is the majority of Civitai, judging by the postings.
❤️AGREE!