REV Mix版本v1.2 (ID: 6361)

Update 2:

ReV_3 released. Read model version description for more info

Update

ReV_2 released. Read model version description for more info

Intro

This model is mainly intended for Portraits and Full Body (Depth of Field) like semi-photorealistic pictures. This can do landscape however, not well. I like this model, jack of all trades, master of none. No restrictions on use of this model, just credit to me and other model creators used in this merge.

Workflow

minimal workflow logic based on Protogen's: Reddit

(Mainly used Difference sum from SD1.5-pruned as tertiary model)

Recommends

Resolution:

• 2:3 ratio (i.e 512x768)

• 1:1 ratio (i.e 512x512 or 576x576)

Best Samplers:

• Euler a (Recommended)

• DPM++ 2M Karras

• DPM++ SDE Karras

• PLMS

CFG scale:

• Portraits: 7-9

• Landscape: 9-12

Steps:

• Portraits: 30-40

• Landscape: 40-100

Hires.fix:

• Upscaler: ESRGAN_4x or ESRGAN_4x+

• Upscale by: 1.5-2

• Denoising strength: 0.3-0.5 (0.3 used in samples)

• Clip: 1

For best results use the: Standard SD VAE

bad-image-prompt-v2

SFW: put 'NSFW' and 'uncensored' in negative prompt, SFW in regular prompt

NSFW: put 'SFW' and 'censored' in negative prompt, NSFW in regular prompt

Less negative prompts the better

License

creativeml-openrail-m

描述:

**use Hires.fix with the following (or similar) settings:

• Upscaler: ESRGAN_4x or ESRGAN_4x+

• Upscale by: 1.5-2

• Denoising strength: 0.3-0.5

Note:
From scratch, More Realistic version compared to V1, strictly for people models, landscape style generations will not come out well.

Major models used

hassanblend-1512

URPM v1.2

RFKTR's Darkdream Diffusion

F222

训练词语: