<h2 id="ken-takakura-(okarun)-or-lora-base:-anima">Ken Takakura (Okarun) | LoRA — Base: ANIMA</h2><p>🙏 <strong>Before anything else:</strong> this LoRA took a lot of work — picking the images, building the dataset and tags, training, and testing the checkpoints. If you have a moment, drop a Like and especially post your results in the comments (with the prompt/settings you used) — it helps me a lot to know what to improve in the next versions. 🔥</p><p>⚠️ The model was trained without any NSFW images. Use it respectfully.</p><p>💬 Check out how the prompt examples were set up in the posted images. 💬 Leave your generations in the comments, and tag the LoRA so I can see how it's looking, it would help a lot. </p><p>📦 Trained with mixed images from Dandadan. Releasing the 750 steps checkpoint as it gave the best results — other checkpoints up to 1200 steps available on request.</p><hr /><h2 id="about-this-model">📊 About this model</h2><p>On the vanilla ANIMA base, Okarun comes out too generic — he loses most of his distinctive design, especially his round glasses and messy hair shape, ending up looking like a generic anime boy rather than the character from Dandadan. This LoRA fixes that: it locks in his silhouette, hairstyle, and glasses much more consistently, making him clearly recognizable across different poses and scenes.</p><hr /><h2 id="generation-tips">ℹ️ Generation tips</h2><ul><li><p>On ANIMA base with 12 steps + acceleration LoRA, the output has a hand-drawn animation feel that works great for Okarun's style — highly recommended as a starting point</p></li><li><p>Using more steps or a different ANIMA-based model with quality LoRAs will improve sharpness and detail</p></li><li><p>Higher CFG can help lock in his features more consistently</p></li><li><p>More steps on the base model are possible but change the aesthetic — personal preference applies here</p></li><li><p>Check the prompt examples used in the posted preview images — they show how to structure tags for each scene</p></li></ul><hr /><h2 id="checkpoints">📊 Checkpoints</h2><p>StepsResult⭐ 750 Best balance of style and likeness — recommended, the one being released</p><p>Up to 1200 Available on request — slightly more "locked in" details, but composition/pose gets more repetitive past 750-1000</p><p>Want a different checkpoint? Just ask in the comments — I have steps available up to 1200 and can upload them on request.</p><hr /><h2 id="generation-settings-used-for-preview-images">⚙️ Generation settings used for preview images</h2><p>All preview images were generated using:</p><ul><li><p>Base model: Anima 1.0</p></li><li><p>Acceleration LoRA: strength 1.0</p></li><li><p>CLIP: anima default</p></li><li><p>Steps: 12</p></li><li><p>CFG: 1.0</p></li><li><p>Sampler: <code>er_sde - beta</code></p></li></ul><hr /><h2 id="one-more-thing">👍 One more thing</h2><p>I'm just getting started with this — if you have a moment, drop a Like and share your generations with the settings you used. It really helps me improve the next versions. Have fun! 🔥</p>

Files with SHA256: a4c8ad97e0b2165b906dacb2c46fcff47ca990a412fb1bb2dc130eea1a1495ce