4thLyco merges

All credit goes to @Minthybasis for his finetune and the extract.I simply merged the latter into my models and found that it learns more tags, improves the quality.It only works right if you use this node, with these settings.Basically, the data contained within the extract does not overwrite Pony's existing data related to anatomy and composition, instead it adds to it's style blocks. Using a normal "Load Lora" node will not have any effect over the style, since it's spreading the data over the entire model.As a result of this, more tags from danbooru like artists, characters and general concepts will be detected, changing the image without touching the overall composition (where characters and their limbs are).It should work with any pony model.Here is a study of the artists contained in 4thTail, which the extract will add into the model and be able to recognize. Some are less pronounced than others, but it works fairly well.This is a manual study so there may be more.Characters sorted alphabetically.Although, I did not merge it into Insomnie because it gets rid of the dithering. If a model has special training data like that, it'll overwrite it (it goes over the style blocks).It (most likely) won't work on realistic models, since the training data isn't focused on anime/furry artwork.