/upscale
solo, male, mature male, (male anthro dragon:1.3), (yellow body:1.1), (standing:1.3), (kemono:1.2), (starry sky:1.23), t-shirt, (clothed), (looking up:1.4), (raised head), (side view), galaxy, aurora, (rainbow:1.3), detailed eyes, portrait, dragon tail, horn, (detailed eyes), (outside:1.35), grass, hill, shooting star, (dark:1.4), (particles ,firefly, blue glowing:1.3), detailed background, 8k hd, (dark shadows, wide dynamic range, hdr, low light:1.2), by Pino Daeni, barazoku, [by dagasi, by zixiong], cygames, by null-ghost
indigowing
This is SE02_vpred, a similar v-prediction model to SE01 with high color saturation and contrast, it seems to have more details than SE01, but the generated backgrounds seem to be more random.
Remember to download .yaml config file and place it alongside the model files, rename the config file name to be the same as the model name.
Using different textual inversions (embeddings) may significantly affect the outputs.
Use CFG rescale extension plz, with a value of 0-0.5(but I think it’s ok to not use it xd)
Clip skip = 1.
中文:
这是SE02_vpred,和se01比较相似的一个vpred模型,有着高对比度和饱和度,与se01相比似乎有更多细节,但是生成的背景更没有逻辑性。
记得下载后缀为.yaml的配置文件,重命名为和下载的模型一样的文件名,并放在下载模型的同一个路径。
使用不同的负文本反转似乎对输出影响很大。
可用可不用CFG rescale插件。
clip跳过为1。