This LoRA captures the illustration style of vazet, a Pixiv artist known for soft painting, delicate linework, and warm color tones.

Trained on a curated dataset of ~132 images, this LoRA works best with SD 1.5 / SDXL anime base models.

Trigger word: vazetstyle

Recommended settings:

- Weight: 0.6 – 1.0

- Base model: Any anime-style SD 1.5 / SDXL checkpoint

- Steps: 20-30

- CFG: 4-5

Works well for: female characters, soft lighting, semi-realistic anime style.

If you like this LoRA, consider supporting my original creations on Afdian / Pixiv Fanbox.