PhotoMerge - CivArchive (CivitAI Archive)

PhotoMerge - v1.0-inpainting

works well with natural language with some tags at the end. use updated v1.5 vae (or whatever vae you want, i'm not your mom).

trigger words aren't need imo. I also recommend a low cfg value like 3.5 or 5. 512x768, 512x512, or 768x768 likely works best, then use hi-res or img2img to get a higher resolution. model was not created with nudity in mind, but it can be unintentionally horny sometimes, so prompt accordingly (simpler prompts get hit harder by this, so be descriptive in what you want).

images generated with DEIS using onnxdiffusersui with max attention slicing, so you probably won't be able to recreate them with auto1111 due to how it handles seeds. quality should be comparable though.

Description

inpainted the eyes and mouth as an example. not sure how strong inpainting is.

FAQ

Comments (3)

saftleFeb 7, 2023· 2 reactions

CivitAI

Looks great! Which models did you merge with? Anything with dreamlike or seek.art (including protogen)?

xzuyn

Author

Feb 8, 2023· 1 reaction

I grabbed all the models that looked the most realistic in my eyes and merged them at 50% in groups then merged those at 50%. results aren't that great but it is my first time merging.

I probably have to add a license or whatever for using dreamlike, but the merge is;

Dreamlike-Photoreal_v2-0_RealisticVision_v1-3_5050 = (Dreamlike-Photoreal_v2-0 + RealisticVision_v1-3 @ 50%/50%)

PortraitPlus_v1-0_AnalogDiffusion_v1-0_5050 = (PortraitPlus_v1-0 + AnalogDiffusion_v1-0 @ 50%/50%)

PPAD_ModelShoot_v1-0_5050 = (PortraitPlus_v1-0_AnalogDiffusion_v1-0_5050 + ModelShoot_v1-0 @ 50%/50%)

PhotoMerge_v1-0 = (Dreamlike-Photoreal_v2-0_RealisticVision_v1-3_5050 + PPAD_ModelShoot_v1-0_5050 @ 50%/50%)

xzuyn

Author

Feb 8, 2023· 1 reaction

i first wanted to try adding difference to keep the detail from all of them, but it didn't work at all and just generated corruption. i also tried merging 50/50 like ((((ModelShoot_v1-0 + AnalogDiffusion_v1-0) + PortraitPlus_v1-0) + RealisticVision_v1-3) + Dreamlike-Photoreal_v2-0) but the results weren't as good.

edit: i have done this, but using a script I found on reddit to merge 3 models and hopefully keep the strength from all of them. v1.1

Checkpoint

SD 1.5

by xzuyn