Platypus - Photorealism版本v3 Lightning (ID: 407760)

PLEASE LEAVE A REVIEW
____________________________

With Version 3, the unique offering of my checkpoint is variety of unique human subjects (women in particular, but men as well).

Recommended wildcards are as follows:

EX: blackmodel redhead woman

EX: blackmodel blonde man

Race:

• european /caucasian

• latin

• asian

• blackmodel

Look:

• Brunette

• Blonde

• Redhead

Framing:

• Close-up /Headshot

• Upper Body /Mid Shot Photo

• Full Body Shot

Recommendations:
Steps: 30, Sampler: DPM++ 3M SDE Karras, CFG scale: 4
ADetailer for face and hand detailer

Tips:
For hand fix, I use FixHands from https://sliders.baulab.info/weights/xl_sliders/ They're in *.pt format, but NextMeal has posted SafeTensor versions here: https://civitai.com/models/238419/multiple-xl-sliders-age-weight-hands-etc

Version 3 Lightning - 8 Steps - recommended settings:

DPM++ SDE Karras, CFG 1-2, 8 steps

Version 3:

Even better!

Previews are RAW and Unedited, with Dynamic Thresholding, and sometimes ADetailer to detail face and hands.

VAE is embedded

Ingredients:

Version 2 + my Female Nobody Generator + PerfectEyesXL + FixHands + DetailedEyesXLV3

Version 2:

Previews are RAW and Unedited, with Dynamic Thresholding, and sometimes ADetailer to detail face and hands.

VAE is embedded

Ingredients:

Checkpoints:

1. Platypus V1

2. B1 - JuggernautXL8Rundiffusion + CinemaxAlpha1 + CrystalClearXL

3. B2 - ProtoVision XLv6.6.0 + realismarkPlus_realismarkPlus + PhotoPedia XL4.5

4. B3 - RealViz2 + StockPhotosV2 + RealisticFreedom_Wonderland

5. RealWow2

6. SocaBabes1

Version B3:

Rework from V1. Previews are RAW and Unedited, with Dynamic Thresholding, and sometimes ADetailer to detail face.

VAE is embedded

Goals with this model:

• Photorealism, human focus

• Good skin textures

• Contrast

Opportunities for improvement in future releases:

• TBD

Ingredients:

Checkpoints:

1. RealVisXL V2 by SG_161222

2. Realistic Stock Photo v2 by PromptSharingSamaritan

3. RealisticFreedom Wonderland by Lorian

Version 1:

Previews are RAW and Unedited, with Dynamic Thresholding.

VAE is embedded

Goals with this model:

• Photorealism, human focus

• Good hands

• Good skin textures

• Range of nationalities/races

Opportunities for improvement in future releases:

• Better skin textures

• Better medium and full body images

Ingredients:

Checkpoints:

1. JuggernautXL by Kandooai

2. CinemaxAlpha by Viakole

3. CrystalClearXL by WarAnakin

4. PhotopediaXL by C3RBERUS

5. LunaMia by WTFusion

6. DreamShaperXL10 by Lykon

7. SVDN-RealXL by StableDiffusionVN

8. FenrisXL by Freek22

9. Demoncore by Afroman4peace

10. FF Double-X-400-NSFW by Idle

11. ffxl400gb-xl by Idle

12. Omnium by FrenzyX

13. Colossus Project XL by Afroman4peace

14. NewrealityXL by ALIENHAZE

15. AlbedoBaseXL by Albedobond

16. DevlishPhotoRealismSDXL by Devilmarkus

17. CineVisionXL by Socalguitarist

18. ThinkdiffusionXL by ThinkDiffusion

19. NightVisionXL by Socalguitarist

20. SocababesXL by Socalista

21. LomoXL by Ndimensional

22. DucHaiten-RealNSFW-XL by DucHaiten

23. RealVisXLV2 by SG_161222

24. LEOSAM's HelloWorld SDXL Base Model by LEOSAM

25. WyvernMixXL by Wier

26. HassanSDXL by Sdhassan

27. ColorfulXL by recoilme

LORA:

BetterFacesXL by sevenof9247

描述:

训练词语: