🗡️ V3 – Peak in Anime 🗡️
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔥 Refined. Sharp. Powerful.
V3 delivers elite-level quality in anime aesthetics — a model designed for bold visions and clean execution.
💬 Got a great suggestion from
🎯 @iJWiTGS8
📎 Check out his work here:
🔗 https://civarchive.com/user/iJWiTGS8
STEPS: 28
CFG: 3,5
Sampler: Euler a
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
V2
I made a couple of LoRAs and used them to merge the checkpoint. I hope it didn’t make things worse!
The showcase images were created at a resolution of 836x1216 and upscaled to a resolution of 1248x1824 using a second sampler.
V1
Hey! Since I’m planning to train some LoRAs in Illustrious, I’ve merged a checkpoint to start with LoRA merge.
That works for me:
CFG: 6-7
SAMPLER: Euler a (DPM++ 2M Karras also works in some cases)
STEPS: 50
The showcase images have a resolution of 1248x1824
Description
I merged the model with 4 self-made LoRAs.
FAQ
Comments (1)
interesting, seems to at base work as if cfg rescale has been applied
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.