{
    "description": "KenCanMix\n\n\n\nThis merged model is between 2.8D to realistic.\n\nVAE is required: https://huggingface.co/stabilityai/sd-vae-ft-mse-original\n\n\n\nHires upscale + Dynamic Thresholding is Highly Recommended.\n\n\n\nRecommended Parameters*:\n\nSampler: DPM++ SDE Karras, DPM++ 2M Karras\n\nSteps: 25-50\n\nHires upscaler: Latent (bicubic antialiased), ESRGAN_4x, R-ESRGAN 4x+\n\nHires upscale: 1.5+\n\nHires steps: 15+\n\nDenoising strength: 0.4-0.6\n\nCFG scale: 7-10\n\nor\n\nCFG scale: 20-30 + Dynamic Thresholding\n\nMimic scale: 7-10\n\nThreshold percentile: 99+\n\nMimic mode: Half Cosine Up\n\nMimic scale minimum: 3-7\n\nCFG mode: Half Cosine Up\n\nCFG scale minimum: 4-7\n\nClip skip: 2\n\n*This is just how I use so far, feel free to play around and share with us any better one.\n\n\n\n===v.1.0===\n\nIf you want Asian faces, LoRA is MUST, and the weight should be higher, try using [from:to:when] prompts. Or applying embeddings such as Ulzzang-6500\n\n\n\nCredit to the authors of the following:\n\nFantasticmix_Real\n\nLofi\n\nLyriel\n\nChikmix",
    "notes": "v.1.6: Fine tuning based on v.1.5, adjusting lighting, focus and color tone for overexposure or vivid issue.",
    "sd version": "SD1",
    "extensions": {
        "sd_civitai_helper": {
            "version": "1.7.2"
        }
    }
}