example of a good merge using djzMushroomLoveV21 at 0.5
credit: DriftJohnson
These "strong style" Models are intended to be merged with each other and any model for Stable Diffusion 2.1
I recommend merging with 0.5 (50/50 blend) then using prompt weighting to control the Aesthetic gradient.
example merged model prompt with automatic1111:
(wwone:1.2) (yourmodeltoken:0.8)
if you drop the "djz" and the "V21" what remains is the token you need to call up the concept in the model. All examples shown were the Raw Token, no other words. Tokens are case sensitive and in almost all models it will match the filename.
It is possible to merge these models with each other using a different value. It is possible to pair models and then merge those resulting models. In this way we can blend abstract concepts together and then weight the tokens to achieve the result we may wish to create. Of course to eliminate all those tokens, you can simply train a new custom model from the outputs, which means you are back to a single token.
A video explanation will follow, but for now the above explanation should do. We are focused on getting as many style/aesthetic models into artist hands to enhance the creativity already at their finger tips.
Art Freedom for all!!
[all original artwork used for training with full permission from Drift Johnson]
Description
FAQ
Comments (9)
please chill with the spam. it makes it tricky to navigate the homepage. no offense but I had to select the option to hide all your submissions because it's overwhelming everyone else.
Yes, that is how the filters work, please use them.
We work hard and if we drip feed models, a new stable diffusion will be released before the people that want to use them ever get the chance. It's not spam, it's hard work. Thanks for understanding we don't sit on our hands. We have a team of three people making models now - Art is for Humans!
Considering there are very few non-NSFW models on 2.1, it's easy to see what you want to see.
let them do what they do, however - do you really need to make a single page for each model?
there is versioning for this reason - at least you could categorize into concepts like "abstract / human / items / whatever" and put there additional stuff
this way you don't spam (i don't mind personally but i can understand how this could be cumbersome for others)
benefit of that would be to have a nice page with everything and you would have nice discussions collected into one place (to be honest i'm not sure if there are any comments or if someone even tried your models? i also skip it for the most part)
hello Malcolm, yes we do need a page for each model, these are exactly as described. Highly specialized and versatile models designed specifically for merging. Just because we released a lot of models in one go means nothing. The alternative is to spend 3 months drip feeding 2 each day? That's clearly unworkable, there will be a new Stable Diffusion release before anyone can even use them. We have never heard of people complaining about "more free resources" before.
Our definition of spam is very different. We have provided more than 65 Style Models for merging with any 2.1 model or each other. This is done by careful design and a lot of time and thought has gone into this. These models are free and our team are volunteers. Commercial Rights are granted. It's hard for us to see how that might be considered "spam". Textual Inversion cannot do what this method can achieve and you can still use any embeddings you like.
The datasets used for these models are all original and some were used to make the djz SD1.5 Aesthetic Gradients series, so we inherited countless hours of planning and curation from the author. We still have 80,000 images to curate into further models and we are developing more at this current time. An artist would never complain about having more colors in his/her palette.
Our models are harming nobody and benefitting everyone who chose to try them.
If you look at our body of work, you will see we offer instructions & use the tags (allowing filtering). Feedback has been overwhelmingly positive as people are hungry for new aesthetic styles to mix with their own models. It couldn't be easier.
It's easy to just be negative as some people have been, but that is not what the AI Art Space is about. We are here to empower artists with new and interesting tools and that what we continue to do.
TLDR, expect us to do the work. We won't let those people that enjoy the use of our models down. Most importantly, Have Fun while making Art.
You could consider merging a bunch and showing great examples of how they work well. The examples here aren't very good yet. I'd love to see the potential!
@Polygon check the reviews :)
@axsthxticroot I have, but like I said, the examples provided aren't showing great potential.
@Polygon @Polygon @Polygon @Polygon @Polygon cool.
Can someone with 3Gb of Vram do a merge? This with some guro/horror checkpoint will look similar to how russian army doing... except the tech looks way too modern.