Thepit style lora (Anima)

詳細

モデル説明

Lora for the style of the artist Thepit (rather his newest art). Trained on Anima v1.0.

Trigger word: @tr1gg3r

Generation settings recommendation:
sampler: er_sde/normal or euler/sgm_uniform - steps: 28-50 - cfg: 4-5

Recommended lora strength: 0.9 or lower for more flexibility, 0.6 gives already most of the style.

For faster generation, I usually generate at 768*1024, then upscale 1.5x with 4x_ultrasharp and 0.25-0.4 denoising (12-20 steps).

このモデルで生成された画像