<h3 id="2-pass-workflow-for-pony-diffusion-base-model-4h72rksxo">2-Pass Workflow for Pony Diffusion base model</h3>(might work for SDXL models too)Sample images generated with:<ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/559512/ohara">Ohara PDXL v4</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/126343">Better Picture, More Details LoRA</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/402947">Detailed anime style - SDXL pony</a></li></ul>Parameters:<ul><li>upscale latent values are good at range 1.2-1.6 depends on image complexity</li><li>combining two clips also improve details (empty prompt is good too)</li><li>image quality and generation speed is good at 1024x768 resolution</li><li>to preserve checkpoint style sometimes is good to set LoRA clip strenght to 0</li><li>some models works well without score tags at all, which may improve hard concepts</li></ul>Other Features:<ul><li>Controlnet (mistoLine)</li><li>Soft Inpainting (left bool checkbox)</li><li>Img2Img (right bool checkbox)</li><li>FreeU, PAG</li></ul>

Files with SHA256: 0beb3c697b87723330121df882748bb02c9dfad8b28284bf215e2c0f6cd452fc