Make sure to use UI that supports v-prediction, otherwise you will get a noisy blob!See about version for credits!<h3 id="main-parameters:">Main parameters:</h3><ul><li>Steps: 28</li><li>CFG: 4</li><li>Sampler: euler a (euler for v6 and lower)</li><li>Scheduler: normal</li><li>Im also using CFGZeroStar and sometimes Self-Attention Guidance: default on difficult backgrounds)</li></ul><h3 id="positive-prompt:">Positive prompt:</h3><pre><code>number of characters (1girl/1boy), character name, artist prompt, scene/environment/camera angle, action, expression, items, quality prompts (masterpiece, best quality)</code></pre><ul><li>some heavily stylized characters might require prompt: realistic</li><li>you can also try to put nose into faceDetailer</li></ul><h3 id="negative-prompt:">Negative prompt:</h3><pre><code>lowres, worst quality</code></pre><ul><li>url against watermark</li><li>lineart, sketch, dot nose for realism</li></ul>Enjoy..

[BPM] Blink's paradise merge (v-pred)

Revisited thanks to <a rel="ugc" href="https://civitai.com/user/Uberlord">@Uberlord</a>, tamed the models for more prompt adherence, darkness and overall stability. Models used, go like them!<ol><li><a target="_blank" rel="ugc" href="https://huggingface.co/Doctor-Shotgun/NoobAI-XL-Merges/blob/main/noobaiXLNAIXL_vPred10Version-cyberrealistic4-perpendicular.safetensors">Doctor-Shotgun/NoobAI-XL-Merges</a></li><li><a target="_blank" rel="ugc" href="https://civitai.com/models/520661?modelVersionId=1979724">2DN</a></li><li>My LoRAs</li></ol>

Files with SHA256: 337359afa3bc9c06a156d3f305c0950a8f3303fdcf5faa25c04c98be4a22af25