If you do like this work, consider buying me a coffee :)
Updated - Mid Year: Updated with a lora created from a frest dataset merged with ICBINP Final. Should give decent results, but does have the standard AI issues (cleft chin bias etc..)
Recommended settings: DPM++ 3M SDE Karras or DPM++ 2M Karras, 20-30 steps, 2.5-5 CFG (or use Dynamic Thresholding), happiest at 640x960 with a hires fix on top, but if you are happy to hunt through seeds to avoid head duplicates, 768x1152 works quite well also.
Following on from Gorilla With A Brick, I've merged in 10 more photorealistic models at various weights, and some more noise offset to create something that when prompted for photorealism will make you go "I Can't Believe It's Not Photography". It will happily create CGI characters and awesome landscapes as well.
As always, pruned to fp16, the VAE is baked in (SD-v2 840000 - ignore the VAE requirement above as the system is borked!), and the man in the suit can be used as the calibration prompt on any UI!
Description
Added Analog Madness, Am I Real, and TangBohu to the mix. Then merged in the totally awesome add_detail lora
FAQ
Comments (12)
This is by far the most incredible model iv ever stumbled across!! the details and it knows characters!! honestly if you have a donation page id proudly throw you a coffee/tea for your hard work with this, i am mind blown!! easily god tier level of work/effort/love gone into making this model♥
is this model also good for landscape photos with no people portraits?
Ive had some success, theres a couple of examples in the earlier versions. You get the occasional inconsistency (ie water flowing in weird ways) but generally looks realish to me
@residentchiefnz am testing with v5, most are ok and a few are quite good. I will upload a few examples
I can't figure out why this model produces such extremely ugly faces. Even when I remove all prompts related to facial features, or any Loras that might affect the face
I think you should stop updating and making garbages, this model was good at version 30, now getting worse and worse every steps.
Good feedback - what for you is better in v3 that is now worse in v5?
@residentchiefnz Resident, you're a gentleman or gentlewoman, and you should be applauded for handling a rude but potentially helpful comment in the best way possible. Bravo
You could just stay in V3 and it will be your checkpoint to fun with it !
Of about the 30 models I've experimented with so far I think this one is my favorite if not at least top 3. When reusing my prompts and switching around models to experiment I always come back to this one, it just has an all around good quality for a variety of ideas and styles. Even if I generate something in another model I'll bring it back to this with img2img or controlnet and fine tune it here.
why does your model have a triggerword?
Because I have added my own lora into it