Ever noticed how hard was to train the 2.1 model and its derivates? It was because it had a heavily censored text encoder, which is the thing that makes pairs with images and roughly understands what's inside the pictures you provide to the model.
Well, there are better and uncensored text encoders compatible with the 2.0-2.1 models on huggingface, and once you get the model you want to train on in diffusers format, you can competely change the single parts of the model. Now, that's what I have done with this model and I will keep working on it.
I have trained the model on a large amount of women of any kind, using the waifu diffusion tagger, and by doing this the model captured the single features of each woman and mixed them together to get info for the celebrities that the encoder knows about but the unet is not trained on.
Currently the model has low capability on ???? stuff, but I will try to make it work next, until then you can have some acceptable depictions of celebrities, I very well know these photos aren't perfect but it is a huge step forward on the basic model.
Ah, since it is not overfitting, it should be able to deal with other concepts, try yourself and, since it was smartly merged with Illuminutty you can try with their embeddings...
(The smaller file is the Negative Embedding)
This particular version seem to work its very best at 512*512, I mainly trained it at this resolution for the lowest end video cards (it works on my 4G quite fast). Despite that, it generates decent 768*768 images too!
1. Dieses Modell dient nur Lernzwecken. Urheber- und Auslegungsrechte liegen beim Originalautor.
2. Bist du der Originalautor eines Modells, kontaktiere uns bitte zur Authentifizierung über unsere offiziellen Kanäle. Wir schützen die Rechte aller Schöpfer. Hier klicken, um es zu verifizieren.
