Test realistic, photorealistic,
Software lllyasviel/Webui-forges
CKXL-EPS
The base model is CKXL-eps, with trigger words referenced from the base model.
Ob2.0-Vpred
Need more detailed prompts make sure Picture detailed
Use 【day】 to ensure the picture is not too dark
NAI-EPS~Vpred
Quality Tags:
very aesthetic > masterpiece > best quality > good quality > normal quality > worst quality
Negative prompt:
worst quality, bad quality, low quality, jpeg artifacts, old, oldest, bad hands, mutated hands,mammal, anthro, furry, ambiguous form, feral, semi-anthro,
2026.1.1 ###trained picture prompt
trained picture prompt
【3D】【guofeng】【illustration】【toon】【tangfu,chinese traditional hanfu,qixiong ruqun】【photo-referenced】
Reg picture prompt
【1photo】【1woman32】【7788RE】【7788WH】【ALAN】【FENGJING】【HAILUOKE】【LINGXIAOQIAN】【NIGHTYY】【XIAOZHUANGFOX】【travel】
you can use 【SDXL1.0_VAE】+ 【t5xxl_fp16】 or 【t5xxl_uncharted-fp16】【refiners/sdxl.text_encoder】【refiners/sdxl.juggernaut.v10.text_encoder】【lopho/Wan2.2-T2V-A14B-Diffusers_bf16_text_encoder】【PixArt-XL-2-1024-MS-BF16-TEXT_ENCODER】【mistral_3_small_flux2_bf16/Comfy-Org】
and use 👇👇👇👇👇
Description
CKXL-EPS0.2+ RE183MG0018 LORA + 5.3FHD
Added training images including 【illustration】, 【guofeng】，【3d】，【toon,toon \(style\)】，【traditional chinese hanfu, tangfu, qixiong ruqun】
Switched to a new AI modelling. Train at a resolution of 1248x1248 using 4Kimages.
换了新的模特。以4K的图片 训练1248x1248的分辨率训练。