Lora trained in recreate Tianliang Duohe Fangdongye's artstyle. Just a little name...
Anima Preview v2: LoKr
I was able to train a lokr and it seems to be better for styles than regular loras (as they was on SD models). I tested them on Comfy
Weight: .8 - 1
Anima Preview v1: LoRa
I was playing with this model since their release. It's a pretty cool improvement for those models that requires a lot of resources for run. It feels like a illustrious v0.1 it needs several tricks to make it work, but using any lora, it gets really easy to use.
This was a test of train a Lora. I selected this one due the native style mess the hands rlly hard, and It works pretty well without any special config.
Try use it in a tag + natural language mix. Isn't consistent like is z-image but it is easier to use than Neta, more powerful than SDXL, uses almost the same vram than SDXL (lower if you use quantized model) . Is Slower than SDXL but it doesn't burn your computer like Flux or Auraflow does.
Weight: 1 - 1.2
Pony v1:
This one is maybe a little overtrained, try use a lower weight than 1 if you have problems.
Weight: .8 - .9
Description
FAQ
Comments (4)
Love all of your loras! Any chance you might make an illustrious version of this style as well? Or are you moving on from Illustrious? (I hope not!)
I'll continue doing Illustrious and Pony loras under request (like yours). But I highly recommend give it a try to this model, as any other, it have their own issues and features. But in general, if you know how to use Illustrious, you know how to use Anima. In general, is a Noob vpred easy to use.
Any pointers to where/how to train an Anima LoRA?
I trained it using diffusion-pipe, a project made by the same creator, but you need Linux to run it. About training config, actually I used the same SDXL params and dataset/tagging and it works. It didn't get crazy like z-image or flux (those that you need no more than 20 images and forcing you to natural lang). I need do some test more before get crazy doing Loras, but in general, it's pretty easy to train. (if you already know train sd models) .