This model depicts matcha, the mascot of the discord server Cafe dot

This is one of my first custom LoRA's, and the first one i ever published.

There are no training tags, but you will get the best results with:

20-40 steps

10-14 cfg scale

512-756x512-756 resolution

Clip skip: 2

Model: anyloraCheckpoint_novaeFp16.safetensors

VAE: clear VAE

And from some trial and error, these prompts work best:

Positive: "sharp, high quality, smooth, clean, anime style, cat, green fur"

Negative: "low quality, blurry, traditional art, pencil, realism, realistic, trademark, wrong anatomy, error, compression"