This LoRa generates Ni Sha from the donghua "War Against Heaven".

Trigger words: Ni Sha

on my own behalf, I can add the following:

the permanent initial construction of building a prompt for generating this character should be exactly like this: ...<lora:Pony Lora - Ni Sha -- War Against Heaven:1>, Ni Sha, forehead mark, ((long hair, short hair, ponytail))...

Regarding the mark on her forehead: I do not know if it is being removed, because I tried to follow only the original during image generation

It took 59 images to train LoRa

Weight: 1.0 - 1.5

1.0 - this value is suitable for you if you do not want to achieve 100% accuracy and want to experiment with other styles, locations or clothes

1.25 - this value will be useful for balancing: the relative accuracy of face and clothing generation ... be careful, because this value is poorly amenable to manipulation with clothes

1.5 - allows you to generate an almost accurate image of a given character ... almost no manipulation of the text will allow you to change her clothes and emotions, too, it becomes difficult to control, most likely, you will have to lift the weights of each necessary parameter

LoRa corresponds to the original only if the CFG scale = 4.5 (I haven't tested it on other values) - it makes sense to play with this value ... I may have missed a more stable value

I will be very glad to receive your feedback and comments, as well as published images and I will be very grateful if you can spread this LoRa mentioning me

During the creation of LoRa, I wondered: "why is this character still not in the AI space, because this Chinese animated series was released in 2023?" ... This time, you can consider that I really was the very first to create this character for the AI space

For those who want to create this character on their own, I left a 7z archive in the DataSet archive, which contains 310 images that I made on my own, since there are very few images of this character on the Internet, somehow it goes beyond the number 20