v.2 -- extended anima range of color, structure, prompt adherence, etc..
-- attention --- this model extends the bases knowledge bringing life to them
-what that means is- this model will do 3d, pixel, 2d, flat, glossy, realistic, realism, you name it, its in there
Chiyo can handle crazy prompts, complex prompts, of course a typical proper formatted anima prompt, Chiyo is a blast to use, it's ability to try and use all tags and create a coherent image is fun to test and play with : )
i dumped a ton of sample pics, a few the same as posted here, on my anime site and my creative image site::
anime related https://animemashup.com/chiyo-anima-v2-sample-gallery.html
creative related https://mindblownart.com/chiyo-anima-v2-model-sample-gallery.html
If you want to look over 150 images fast to see what this model can do, check the gallery of animemashup.com or mindblowart.com
additional samples below -- you can test it online, today after bidding closes
Chiyo, a name steeped in Japanese heritage, offers a unique blend of elegance and profound meaning. Evoking "thousand generations," it carries a beautiful wish for longevity and auspiciousness, a sentiment that resonates deeply through its classic yet contemporary appeal. This is a name for parents who appreciate a subtle nod to tradition and a deep, hopeful meaning, without sacrificing a light, graceful sound. Unlike many names with historical roots, Chiyo manages to feel both ancient and utterly fresh, perhaps due to its gentle phonetics and the enduring significance of its meaning in Japanese culture. It’s a choice that speaks to a desire for enduring beauty and good fortune for a child.
i fused my lora to this model until i can do a proper training which is trained so i am labeling this as trained, since the thing affecting the outcome is my sole effort.
Description
extended anima range of color, structure, prompt adherence, etc..
FAQ
Comments (3)
i bid the version 2 to be online this week
there was an issue.. and waiting for them to fix issue so it can be used online
You model looks promising tho most non dedicated style models (ie a model dedicated to a flat color anime style or a model dedicated to a realistic style, etc...) tends to struggle with style consistency and force you to go crazy on style related tag scores (ie: flat color, anime coloring, western comics (style), realistic, photorealistic, etc... set to very high scores). Except that doing so reduce the weight of other tags inside your prompt. So this isn't good. So how consistent is your model when using a style tag when generating a portrait, then a upper body shot, then a full body shot, then a landscape focused shot? Will it keeps the style tag that is used or ll it switch to a different style at random or dillute the style tag?
i was just looking for list or text files with the model knowledge to test that stuff : )
i want to build a workflow that can grab tags from folders. I know there are a few sites with the info, but where did they get the original info. circlestone has artist tags in a list, i could not find the rest.
basically i learned alot working with morph challenges --materials
I teach models what they are missing i fill in gags as a visual artist troubleshooting a car (model) that needs a tune -up (finetune)
i will be able to bring anima pretty far, it is an excellent base -this is just a start