兩個版本
都是權重越大越偏向寫意畫風,在1左右是工筆畫風
V1 由圖像訓練而來權重不要超過1.5 適合權重在0.7~1.3左右就夠了,如果令牌太長可以適度增加權重
V2 由V1 和一些過往訓練的LORA混何而來權重最大2.0 適合權重在1~1.75,如果令牌太長可以適度增加權重
主要是畫景色或者靜物
如果要畫人,建議V2權重小於1
Two versions
The larger the weight, the more favourable the style of painting, at around 1 is the style of brushwork.
V1 by the image training from the weight should not exceed 1.5 suitable weight in 0.7 ~ 1.3 or so is enough, if the token is too long can increase the weight appropriately.
V2 from V1 and some past training of LORA mixed from the maximum weight of 2.0 suitable for weight 1 ~ 1.75, if the token is too long you can increase the weight appropriately
Mainly for landscapes or still lifes, if you want to draw people, we recommend that V2 weights are less than 1.
This version adds a greater proportion of inksplashing and white negative space
1. The rights to reposted models belong to original creators.
2. Original creators should contact SeaArt.AI staff through official channels to claim their models. We are committed to protecting every creator's rights. Click to Claim
