I prefer the flat style, so I integrated my own Lora and created this flat big model, which may continue to be iteratively optimized in the future.

Update instructions:

The v1 version has better recognition of prompts, can express characters/scenes/objects, and can be used as

operational illustrations.

The beta version, the original style is sharper, but the compatibility with the prompts is insufficient.

(When adding words such as tree, mountain, plant, leaves, etc., it may become a pure landscape, and other objects are not easy to come out)

Choose according to your preferences!

-------------------------------------------------------

No external VAE required, Or use animevae, (8400vae might be a little blurry),

sample: Euler a,

CFG: 7-8,

HD-fix: R-ESRGAN 4X+Anime6, redraw amplitude of about 0.3

It is suitable for drawing up multiple pictures at once and selecting the ones you like. Have fun drawing cards!!