>>34596694You mean WD 1.2? 1.2 is the current one with 1.3 releasing in three days
It doesn't know many characters because of the low number of training images compared to NAI's.
You can try to make textual inversion embeddings or dreambooth models on the characters you want, which is what I did
