VERISIMILANIMA

AVAILABLE FOR FREE GENERATIONS @ TENSOR ART

~~__________________________________________________________________________________________________~~

🔷ABOUT THIS MODEL:

TLDR: Experimental Surgical Merge of several ANIMA models to create a model based on Anima 1.0 with added characters and improved Realism style.

I present the Anima version of my model mergers, Verisimilanima, a wordplay of Verisimilitude that means "the appearance of being true or real" and ANIMA.

My goal was to add new character data to ANIMA Base 1.0 and improve the realism, my only option atm is to merge models but for my goal I had to create new ComfyUI Nodes.

These nodes allow me to control and merge specific layers (cross attn, self attn, mlp, adaln cross attn, adaln self attn and adaln mlp) in every of the 28 blocks in the Anima model. With this kind of control, I was able to hand pick what to merge. This method is the same concept I used in Verisimillustrious models but focused on SDXL.

A lot of trial and error happen and one hard drive died, so any support is appreciated.

~~__________________________________________________________________________________________________~~

🔷MODELS USED:

(Please check these amazing models, without their hard work this merge wouldn't be possible)

Kirazuri (Anima): For the new character data, but avoided any style related changes.

RealCosplay&Anime&3D ALL-IN-ONE: For the improved realism with minimal style modifier.

ANIMA 1,0 + (KIRAZURI - ANIMA 1.0) = MERGE A

MERGE A + (REALCOSPLAY&ANIME - ANIMA 1.0) = MERGE B

MERGE B + (KIRAZURI - ANIMA 1.0) (2nd pass) = FINAL MERGE

Note: Merger was block and layer specific every time, no full model mergers were used.

~~__________________________________________________________________________________________________~~

🔷HOW TO USE:

Follow the typical NoobAi/Illustrious/Anima prompt structure first.

#gender, character, series, character description, artist tag or style, quality tags, camera tags, theme or background tags, actions, clothing.

Then add natural language describing the style, but also if you need specific details for the character.

Check the EXAMPLE IMAGERS for prompts ideas, I will try to keep them simple to allow a quick start.

~~__________________________________________________________________________________________________~~

💠PERMISSIONS AND RESTRICTIONS:

✔Use only on TensorArt/CivitAI and for free Generation at TensorArt

✔Mergers allowed (Just credit and link model)

✔Share generations

✔Credit me or name the model when used in any non-generation capacity

⚠Availability outside of the designated websites is my prerogative, will make it known in the description for official places where it would be available.

❌Use this Model or Merges of this model outside of CivitAI or TensorArt/Hub

❌Sell Art

❌Sell Model

❌Claim it's your Art

~~__________________________________________________________________________________________________~~

REGARDING AI: My stance is that AI doesn't create/produce real art, but it allows to materialize concepts/ideas, but I don't consider it art. So, IF I include prompts in the examples, I will most likely not include artist tags in the prompts. It's my way to respect all the awesome artists whose art was trained on, so please understand.

~~__________________________________________________________________________________________________~~

License

ANIMA base 1.0 License: CircleStone Labs Non-Commercial License.

Built on NVIDIA Cosmos.

Verisimilanima - βeta