Эта модель — мой первый тест, не стоит слишком надеяться.😳

v1.1 and v1.06 Trained on Pony Diffusion V6 XL, CFG scale is about 5.

v1.06 should be the best one now.

Amazingly, you can use v1.06 and v1.1 versions at the same time, and they can work normally.As long as you make sure that their strength adds up to less than 1.

The characteristics of this lora model are: the (back spikes) of characters mainly (dragons) and the (spiked tail) are well generated.But there will also be (spikes) on the (large pectorals). ACTUALLY, I like it.Maybe you don't like it, you can try it in negative prompt.

Because of the training image, (three-quarter view, full-length portrait) will be better generated.

Moreover, this model has a very strong depiction of feet, especially (toes) and (large pectorals), and (featureless croth), which comes from the painting style of the artist.

Oh, the (horns) of characters are not well generated. :(

prompt:

valstyletest1, anthro, three-quarter view, full-length portrait, featureless croth, colored eyes, colored skin, muscular, large pectorals, toes

I prefer to add:

dragon, sharp teeth, open mouth, claws, glowing spikes, back spikes, spikes on body, spiked tail

(glowing spikes, glowing striped body) is very good for (black skin)!!!

negative prompt (If you want to stay away from the comic style):

(comic style:1.3),(sketch:1.4)

I'm not sure if this will work.

Maybe you can try to use it with other artist style models↓.