HighRiseMixV2 / README.md
0RisingStar0's picture
Update README.md
2f8b6b7
|
raw
history blame
2.97 kB
metadata
license: creativeml-openrail-m
tags:
  - stable-diffusion
  - stable-diffusion-diffusers
  - text-to-image
  - diffusers
HighRiseMixV2.5

HighRiseMixV2

U-Net mixed model specialized for city and skyscrapers background.

FP16 Pruned version(No EMA).

(Quality change may occur in very small details on buildings' textures)

V2 Update Log :

Added models : AikimixPv1.0, Counterfeit V2.0, pastelmix-better-vae

Adjusted character style(more cute, anime style)

V2.5 Update Log :

Added models : AikimixCv1.5

Just some very little changes to textures adjusted to my taste. It doesn't matter which one to use. There are pros and cons between V2 and V2.5 so you can just use what you want.

Recommended prompts :

(masterpiece, best quality, excellent quality), ((1girl, solo)), sky, city, (skyscrapers), trees, pavement, lens flare

EasyNegative, moss, phone, man, pedestrians, extras, border, outside border, white border

(EasyNegative is a negative embedding : https://huggingface.co/datasets/gsdf/EasyNegative)

Recommended settings :

Sampler : DPM++ 2M Karras OR DPM++ SDE Karras

Sampling steps : 25 ~ 30

Resolution : 512x768 OR 768x512

CFG Scale : 9

Upscale is a must-do!! Otherwise, you won't get intended results.

Upscaler : Latent (nearest)

Hires steps : 0

Denoise : 0.6

Upscale 2x

Recommended VAEs :

kl-f8-anime2

orangemix.vae.pt

Mixed models :

AbyssOrangeMix2_NSFW, AnythingV4.5, AikimiXPv1.0, BasilMixFixed, Counterfeit V2.0, CounterfeitV2.5, EerieOrangeMix2, pastelmix-better-vae, PowercolorV2

(Thanks to everyone who made above models!)

This is my first mixed model being uploaded to public site, so feel free to give feedbacks as you wish, I'll try and work around with it.