<h1 id="raemu-xl-km7h5eq59">Raemu XL</h1>Raemu XL is a SDXL-based anime model designed for 2.5D anime-style generation. It excels in producing high-quality, semi-realistic anime visuals with rich textures and refined shading.<h3 id="v5-5nv7lsfxi">V5</h3>Raemu XL V5 is a finetuned iteration based on <a rel="ugc" href="https://civarchive.com/models/1188071/animagine-xl-40">Animagine 4.0</a>, further enhancing its ability to generate aesthetically refined 2.5D textures. This version improves lighting, depth, and overall visual quality while maintaining the signature anime look.Ideal for high-quality illustrations, character art, and stylized anime rendering, Raemu XL delivers a balance between artistic expressiveness and fine detail.<hr />Why Early Access?Early access helps cover the cost of training and development. By purchasing early access, you're directly supporting the development of my models. If you'd like to support me further, you can also <a target="_blank" rel="ugc" href="https://ko-fi.com/raelina">buy me a coffee on Ko-fi!</a> Your support keeps the project going and ensures even better models in the future.<hr /><h3 id="how-to-prompt.-d4s220nzh">How to prompt.?</h3>read Anim4gine GuideLines <a target="_blank" rel="ugc" href="https://cagliostrolab.net/posts/optimizing-animagine-xl-40-in-depth-guideline-and-update">here</a><h3 id="recommended-prompts-p3z2bjdn7">Recommended Prompts</h3>To guide the model towards generating high-aesthetic images, use negative prompts like:<pre><code>lowres, low score, bad score, worst quality, bad anatomy, bad hands, sketch, jpeg artifacts, text, error, signature, watermark</code></pre>For higher-quality results, add quality at the end of your prompts like:<pre><code>masterpiece, high score, great score</code></pre><h3 id="recommended-settings-l0l17shew">Recommended Settings</h3><ul><li>Sampler : Euler a</li><li>Sampling steps : 28</li><li>CFG : 6</li></ul><ul><li>Upscaler : <a target="_blank" rel="ugc" href="https://nmkd.de/?esrgan">4x-NMKD_YandereNeoXL</a></li><li>Hires step : 10-15</li><li>Denoising Hires : 0.1-0.3</li></ul><h2 id="license-fbjvvmf68">License</h2>V5 model adopts the original <a target="_blank" rel="ugc" href="https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/blob/main/LICENSE.md">CreativeML Open RAIL++-M License</a> from Stability AI without any modifications or additional restrictions. The license terms remain exactly as specified in the original SDXL license, which includes:<ul><li>✅ Permitted: Commercial use, modifications, distributions, private use</li><li>❌ Prohibited: Illegal activities, harmful content generation, discrimination, exploitation</li><li>⚠️ Requirements: Include license copy, state changes, preserve notices</li><li>📝 Warranty: Provided "AS IS" without warranties</li></ul>Please refer to the <a target="_blank" rel="ugc" href="https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/blob/main/LICENSE.md">original SDXL license</a> for the complete and authoritative terms and conditions.<h2 id="below-are-for-legacy-v1-v4-j1xly4ea0">Below are for legacy V1-V4</h2><hr />SDXL anime base model that focused in 2.5D Anime. RaemuXL can generate high-quality anime images. This model merged from <a target="_blank" rel="ugc" href="https://huggingface.co/cagliostrolab/animagine-xl-3.1">Animagine XL 3.0/3.1</a> as a base. Compatible with any Lora that trained from Animagine XL 3.0/3.1.<h3 id="how-to-prompt.-31g8sb8ue">How to prompt.?</h3>read Animagine GuideLines <a target="_blank" rel="ugc" href="https://huggingface.co/cagliostrolab/animagine-xl-3.1#usage-guidelines">here</a><h3 id="recommended-settings-c405o9cox">Recommended Settings</h3><ul><li>Sampler : Euler a</li><li>Sampling steps : 25-35</li><li>CFG : 7</li></ul><ul><li>Upscaler : <a target="_blank" rel="ugc" href="https://nmkd.de/?esrgan">4x-NMKD_YandereNeoXL</a></li><li>Hires step : 10-15</li><li>Denoising : 0.1-0.3</li></ul><h3 id="recommended-settings-for-lightning-version-g22em5pni">Recommended Settings for Lightning version</h3>The Lightning version has faster generation with a slightly decreased quality<ul><li>Sampler : Euler a</li><li>Sampling steps : 8-10</li><li>CFG: 2-2.5</li><li>Upscaler : <a target="_blank" rel="ugc" href="https://nmkd.de/?esrgan">4x-NMKD_YandereNeoXL</a></li><li>Hires step : 4-6</li><li>Denoising : 0.1-0.3</li></ul><h3 id="recommended-prompts-1b9qj5js8">Recommended Prompts</h3>To guide the model towards generating high-aesthetic images, use negative prompts like:<pre><code>(low quality, worst quality:1.2), very displeasing, ugly, poorly drawn, signature, watermark,</code></pre>For higher quality outcomes, prepend prompts with:<pre><code>masterpiece, best quality, very aesthetic, absurdres,</code></pre><h3 id="multi-aspect-resolution-ibeqo78jn">Multi Aspect Resolution</h3>This model supports generating images at the following dimensions:| Dimensions | Aspect Ratio ||--------------|---------------|| <code>1024 x 1024</code> | 1:1 Square || <code>1152 x 896</code> | 9:7 || <code>896 x 1152</code> | 7:9 || <code>1216 x 832</code> | 19:13 || <code>832 x 1216</code> | 13:19 || <code>1344 x 768</code> | 7:4 Horizontal || <code>768 x 1344</code> | 4:7 Vertical || <code>1536 x 640</code> | 12:5 Horizontal || <code>640 x 1536</code> | 5:12 Vertical |<h3 id="-8qdggxw0r"></h3><h3 id="license-m4mk2sghb">License</h3>Raemu XL now uses the <a target="_blank" rel="ugc" href="https://freedevproject.org/faipl-1.0-sd/">Fair AI Public License 1.0-SD</a> inherited from Animagine XL 3.0, compatible with Stable Diffusion models. Key points:1. Modification Sharing: If you modify Raemu XL, you must share both your changes and the original license.2. Source Code Accessibility: If your modified version is network-accessible, provide a way (like a download link) for others to get the source code. This applies to derived models too.3. Distribution Terms: Any distribution must be under this license or another with similar rules.4. Compliance: Non-compliance must be fixed within 30 days to avoid license termination, emphasizing transparency and adherence to open-source values.<h3 id="-8j2wzs6rx"></h3>For merge recipe go to my <a target="_blank" rel="ugc" href="https://huggingface.co/Raelina/Raemu-XL-V4#merge-parameter">Huggingface Repo</a>

Raemu XL

I tried my hand on the Lightning version.I merged with Lightning LoRA and finetuning it after with 860 high-quality images. What are the differences with v3 using Lightning LoRA:<ul><li>More improved quality (I did finetuning on the Lightning checkpoint)</li><li>Reduced noise and artifacts</li><li>Reduced reddish coloring</li></ul>Recommended Lightning Settings:<ul><li>Sampler: Euler a</li><li>Sampling steps: 8-10</li><li>CFG: 2-2.5</li><li>Hires step: 4-6</li><li>Denoising: 0.1-0.3</li></ul>Feel free to experiment, but I still recommend the settings above.

Files with SHA256: bbd321d4a3a7d93196b32b619f74574db8bb2e342f8a8355eca21e5a82c56dea