This LoRA is designed to generate soft, charming, anime-inspired furry females across a broad range of species. It emphasizes expressive eyes, clean linework, gentle color palettes, and appealing character silhouettes while maintaining flexibility for both fantasy and slice-of-life themes. The model supports a full spectrum of anthro and kemono designs without locking you into a specific character.Recommendations:<ul><li>1024x1024 resolutions (SDXL), can work up to 1536x1536, image breaks apart at 2048x2048 or above</li><li>LoRA weights between 0.6 and 1.0</li><li>Prompting like this: [Natural Language (Character descriptions + Clothing + Environment)] + [Tags] - especially recommend the tags anime, digital painting, detailed, furry, anthropomorphic, female.</li><li>Example prompt: "<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">A small, delicate female anthropomorphic cat with a slender graceful frame, narrow shoulders, thin limbs, tiny feet, long red hair cascading over her shoulders, vibrant flame-point bicolor fur with creamy white chest and belly and bright orange back, legs, and tail, small bust, dainty hands with retractable claws, tapered ears, whisker pads, vertical slit pupils, thick eyelashes, and soft plush fur with subtle directional flow, A long robe with wide sleeves and a sash belt, featuring a flowing mystical design suitable for a wizard or healer, A stormy cliffside overlooking a turbulent ocean, with lightning striking jagged rocks and waves crashing below. fantasy, female, furry, anthropomorphic, full body, digital painting, detailed, soft lighting, anime."</li><li><span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">For ZIT: Z-Image Turbo V. 1.0, 9 Steps, 1 CFG, res multistep, simple</li><li><span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">For ILL: Illustrious V. 1.0, 30 Steps, Euler a, normal, CFG 3.5 (with 4.0 PAG) / 7.5 (without)</li></ul>With time, the dataset will grow and eventually this aims to become a checkpoint itself. The dataset is based on 4k images SFW I generated myself - I uploaded some of them here on my page, you can check the generations methods and ressources there.

KemoKoko

I was curious how the dataset would be applied to a Illustrious model and I'm also liking the results so far. Edit: Changed the file a to an earlier epoch

Files with SHA256: 0ac63c82976a10baa213d1dd53cc672bb399477ca3a4f880530a59619c66d005