I discovered the optimal training parameters in SDXL 1.0. The generated LORA now resembles a real person in real life.Train based on SDXL 1.0Dataset using over 100 carefully selected high quality imagesPlease follow my Facebook <a target="_blank" rel="ugc" href="https://www.facebook.com/AIImagingSD">AI_Imaging </a>Join my discord server <a target="_blank" rel="ugc" href="https://discord.gg/pJ6GGcgAGV">https://discord.gg/pJ6GGcgAGV</a> for free discussionMy Patreon <a target="_blank" rel="ugc" href="https://www.patreon.com/AI_Imaging/membership">https://www.patreon.com/AI_Imaging/membership</a> for exclusive LORAs*************************Best weight is from 0.8 to 1Steps: 40-60, Sampler: DPM++ 2M Karras, CFG scale: 7Model: majicmixRealistic_v6, Denoising strength: 0.3-0.4, Clip skip: 2, Hires upscale: 2, Hires steps: 16-30, Hires upscaler: 4x-UltraSharp******************************************************In our Telegram channel - <a target="_blank" rel="ugc" href="http://t.me/gaidep69">AI Imaging: The new ERA</a> you'll find exciting content related to AI and imaging.We also have a Telegram group <a target="_blank" rel="ugc" href="https://t.me/gaidep6996">Free Group- AI Imaging</a> where you can engage in discussions and connect with like-minded individuals.Additionally, on our YouTube channel - <a target="_blank" rel="ugc" href="https://www.youtube.com/channel/UCQ4LyalfNNNS0U2_ZmEavAA/">AI Imaging Studio</a>, you can explore a wide range of AI-related content.

Files with SHA256: 3a280db95e329b5bfac02de47249317229f0c7c309388da9b8e13787a0bfdb50