Anima 1 base, but realistic in the style.
It produces a mix of realistic styles:
Really gorgeous professional photography
Amateur photos
Semi-realistic drawings
Ugly people
There are probably ways to get more of one type or the other, for now I prefer to keep it flexible, but will aim towards reducing the ugly people outputs.
Description
FAQ
Comments (13)
Im not sure if the outputs are dishonest but using the exact prompts you give with the exact settings provided, the realism and quality is not matched
Are you using the images as workflows in ComfyUI?
just from a quick glance most of the outputs look upscaled. That doesn't mean its impossible
@makiaeveli I'm using a 2-step process. First generate a low-res, then upscale.
It's all in the workflow.
There's no secret or post-processing if that's what OP is talking about.
Run the workflows and you should get exactly the same result.
@6tZ I dont think workflow is saved correctly. Altho I do get rather similar output, just not in terms of image clarity.
EDIT: Figured it out. Issue is you are using outdated/different version of 🔀 Batch Wildcard Sampler, while current version is called 🔀 Batch Wildcard Upscale Sampler.
I cant really figure out your exact settings for upscale, but I got very similar output to yours, so I think its fine.
the outputs are deceptive, theyre not produced natively by the model.
into the trash it goes
Nah, its good model. Just needs to use exactly workflow of author in ComfyUI, nothing else. Good stuff is in that model.
You can use whatever workflow you want.
The workflows are included in the images, use them or don't, you do you.
Nothing is faked, why would I upload deceptive images?
@6tZ I think I can understand the commenter; it would be nice to have a comparation of the model by itself + model with the workflow that includes detailers. That way people don't get the wrong idea on the first view of the images and try your workflow to get the quality of the images. Either way, I think the results of the model by itself aren't bad, tho the workflow does add the extra clarity as the example images. Great model none the less.
@darionk I mean, if someone wants to do that, please go ahead. I agree with that.
I just don't have the time to make such bespoke requests.
The model can output great things if used correctly, and I've shared those.
If people want to figure out other ways of working, please do and share.
There should be plenty of images shared for v1 already that uses other's workflows and shows it working just fine for them too :)
Excuse me. Could you please tell me the correct procedure for using it?
The images contains the workflow used to generate them.
1. Download the image
2. Drag/Drop into ComfyUI
3. Re-link models, get missing nodes (you may need to update some nodes or replace them with a new version, just search the node list).
Or learn how to use Anima from any other workflow :)
Use the settings listed on the model page.
@6tZ Thank you for steps!!