Hello, I wasn't really satisfied with what we currently had on Civitai for Giga, so here you go!Giga style LoRA trained for Anima on 189 non-synthetic images, focused on recreating the distinctive rendering, facial features, body proportions, linework, and color palette of Giga.<h2 id="recommended-settings">Recommended Settings</h2>Base model: Anima Base / Anima-derived checkpoints LoRA weight: 0.8–1.4 CFG: 3–6 Steps: 25–35 Can be use without trigger word.Sampler: er_sde works very well, other Euler-class samplers work well.<hr />If you get good results, feel free to post them on the model page. Feedback is welcome, especially if you notice prompts where the style does not apply well.

Files with SHA256: e0e629618d40112880591ba4836f18eba8cbc3ae8ddb0fcafeb265e0ee21f9ff