V3: Fuller images, booru tags + natural language in captions, overall better colors, more cartoon style. Different age groups and genders. V1 + V2 only trained women,iX comes to Anima!<hr />This uses my trained iX lora/checkpoints from Illustrious mixed in with Anima Base!Works very well with natural language.<hr />YOU NEED:qwen base encoder &amp; qwen image vae<hr />All gallery images should have metadata so you can drag and drop to see the exact things you need along with prompts and settings.<hr />SETUP:Steps: 30 - 50CGF: 4 - 5Sampler: Eular/Eular A (haven't tested others)

iX-Anima

What changed in testing (V2 vs V3, same 26 prompts and seed)<ul><li>Fuller compositions: subjects are larger and more detailed. A dragon over a castle went from a dark silhouette to a bright, detailed red dragon, and a cyclist went from small in the landscape to the main subject.</li><li>Stronger color: boxer, samurai and storm scenes pick up magenta, violet and red light where V2 stayed darker.</li><li>The iX look now holds on men and older people. V2 drifted toward plainer, more realistic faces on them.</li><li>Faces are rounder with bigger eyes, so V3 reads more cartoon than V2. This is the biggest change to the style.</li><li>Pink and magenta show up in more scenes, including ordinary ones like a kitchen or a library.</li><li>Known characters still come from the base model. None of the gallery characters (Yoruichi, Kafka, Ranni, Boa Hancock) are in the training set, and all of them render on-model.</li></ul>Tips and known limits<ul><li>Clean-shaven young men and characters with no age cues can read young, as on V2. Add stubble, a beard, <code>mature male</code> or <code>mature female</code>.</li><li>In wide frames, use <code>full body</code> and say "head down to the feet". A sentence alone can crop the figure.</li><li>Add <code>standing</code> if the pose matters. Some actions still crouch or sit on their own, such as feeding birds or a campfire scene.</li><li>Neon signs can show gibberish, or even the character's name.</li></ul>Recommended settings<ul><li>30 steps, CFG 4, shift 3, euler / simple, about 1–1.3 MP (e.g. 864×1536, 1536×864, 832×1216)</li><li>Text encoder: qwen_3_06b_base. VAE: qwen_image_vae.</li><li>Prompt with booru tags, a plain sentence, or both.</li></ul>

Files with SHA256: d9eebabf851936a3dc5200f9ea88983f985622c288b6e067c38a2166546318f2