SD

placidmix_v1.1

placidmix_v1.1 - AI Model封面圖片
上傳時間
2023年11月19日 下午6:04
使用數
8.9k
評論
滿意  (1)
許可權
  • 允許圖片生成與分享
  • 允許使用者下載此模型
  • 允許生成圖片用於商業用途

推薦 LoRA

描述

https://civitai.com/models/143492?modelVersionId=225483

I'd love to see what kind of results people are getting, so I would appreciate it if you could upload any images that you are particularly fond of.

V1.1

This update is focused on improving the clarity of images that are generated using the mix, while also making colors a bit more vibrant.

V1.1 is generally the more stable version, and as such you will tend to get better results in less generations.

V1.0

This mix came about when I decided to try merging some stylized models together; and even though the results are somewhat generic there is something about them that I just enjoy.

Source Models:

PastelMix

Dark Sushi Mix - Colorful

NyankoMIX

TMND-Mix

Ahnn

Recipe:

v0.1 = PastelMix (0.5) + Dark Sushi Mix Colorful (0.5)

v1.0 = v0.1 (0.7) + NyankoMix (0.3)

V1.1 = V1.0 (0.5) + TMND-Mix_SP-RAINBOW (0.2) + Ahnn (0.3)

Advice:

For recommended parameters there isn't anything special but I've been using:

Sampler: DPM++ 2M Karras

Steps: 30-35

CFG Scale: 7

Clip Skip: 2

It's recommended to use both Hires.fix and ADetailer, and both have been used in the example images

I recommend using 4x-Ultrasharp but other upscalers are probably fine

In terms of VAE you can really use what you want, I've used Vae-ft-mse-840000 in the example images if that is what you are looking for

留言

由 placidmix_v1.1 製作的作品

1