Use danbooru tagging for prompts Do ->not<- use "masterpiece" or "best quality" tags (v9+) Sampler: "Euler a" at / 30 steps Cfg scale: 8 Clip Skip: 2 <---- VERY IMPORTANT 99% of anime models only works with clip skip set to 2 ENSD (Eta noise seed delta) - leave blank kl-f8-anime2 VAE Negatives: (EasyNegative:0.8), (worst quality, low quality:1.2), blurry, bokeh, dof, fog
it’s recommended to use a lower classifier-free guidance (CFG Scale) of around 5-7, sampling steps between 20 and 28, and to use DPM++ 2M Karras as a sampler. But I also tested using Euler Ancestral(Euler A
Sampler - DPM++ 2M Karras (preffered. Personally, I use Euler SGMUniform) Steps - 36+ Clip Skip - 2 Base image size - 512x768 HiresFix - Upscale Factor - 2 Steps - ~15 Denoise strength - 0,6~0,7. Additional - This model can generate good images without much of embeddings. I personally use only badhandv4 or badhand5. You are free to use whatever embeddings
