FaserCore is like a digital artist that can do a bit of everything. Whether you're looking to create stunning landscapes, detailed portraits, or imaginative designs, this model is here to help. It’s versatile and easy to work with, making it perfect for any kind of image you want to create. Think of it as your go-to tool for bringing ideas to life, no matter what they are. And as we release newer versions, FaserCore will keep getting better and smarter, handling even more complex tasks with ease.
Important Usage Tips for SDXL Models
Add Stable_Yogis_PDXL_Positives at the beginning of your prompt section.
Add Stable_Yogis_PDXL_Negatives-neg at the beginning of your negative prompt section.
Workflow for Comfy UI
DMD2 Workflow = https://civarchive.com/models/1215102/sdxldmd2workflowbystableyogi
Pro Version of DMD2 Workflow By Stable Yogi= https://www.patreon.com/posts/121447074
I couldn’t have brought this model to life without the invaluable guidance and mentorship of Stable_yogi. He’s been the ultimate master in this craft, sharing insights and techniques that have taken this model to the next level. If you’re looking to create incredible custom models of your own, Stable_yogi is truly the go-to person.
If you have any suggestions or feedback you can always connect with me on civitai chat.
Description
Thanks for your attention , Your feedback/comments and reviews are much appreciated.
Steps: 20,
CFG scale: 5,
Sampler: DPM++ 2M SDE Karras, DPM2 a, Euler a,
Size: 720x1260,
Face restoration: CodeFormer,
Clip skip: 2
In positive prompt i am using Stable_yogi's Positive embedding - > Stable_Yogis_PDXL_Positives
In the Negative prompt i am using this embedding : Stable_Yogis_PDXL_Negatives-neg
FAQ
Comments (9)
Mega impressive model for first try.
But is it bugged? Some random prompt word make it ultradeformed. Like i used those p1g30n, 1g0r prompts, and while on other checkpoint it adds very good plus detail, on your checkpoint the result is ultradeformed something. And random words act like this too: bus stop(lol). I think its buggy.
Thanks for the feedback, kindly share the images in the gallery below, for me to check. I would love to experiment with those tags.
@Faser_Media totally random prompts, not specific words. When i reloaded Forge, it worked as intended, then i put totally random words, and the result was a mess. Examples: bus stop, result some blurry deformed shit. I sometimes use IMG.0001.JPG prompt to add realism, result was a body horror oil painting. I added those spotify words, i saw in your prompt, result was mess. Then i restarded Forge, and it worked fine. I know Forge can be buggy, but another checkpoint doesnt make this.
@fikejyhe478 thanks a lot for your experiment's, kindly share the images, let me try those keywords too.
@Faser_Media maybe something related to pytorch. Pytorch:2.3.1, python:3.10.6, the newer versions said: its not supported by my card. I have rtx 2070.
@kumameqe525 looks like time to upgrade your Rig !
@kumameqe525 you should definitely be able to use pytorch 2.3.1 on a RTX2070, it seems like the error you're getting has something to do with either your venv or a script that is trying to downgrade your pytorch version. PyTorch 2.X has been working without issue for me... and im on a pascal card... (1000 series LOL)
Hi, what are p1g30n/1g0r promts?
So far its pretty good, steampunk looks very nice. But now and then it messes up background faces. Like they were just sort of stamped on. They didn't blend. Or the hands and feet are gnarled.