<a target="_blank" rel="ugc" href="https://ko-fi.com/diffusiondesign">If you enjoy my models and want to help support my releases, I always need a coffee :)</a>There are only TIs of him available and honestly none of them were good enough for me, so I trained a fresh LoRA and have decided to share. Trained on the 1.5 base at 1024 res.Please don't rip this off as your own, or do anything crazy with it. I am not responsible for how you use this. I am not releasing it for professional or commercial use, this is just for personal use and non-commercial research purposes. Please respect him and please share your results!<ul><li>Use "pedro pascal" as your keyword</li><li>This LoRA stacks well on almost all models I've tried, from photorealistic to anime</li><li>I had to overbake it a bit to get it to learn his characters, so tone it down if it's coming out too "crunchy" or it's only giving you closeups</li><li>Works best between :0.4-0.8</li><li>Using a lower value with a higher keyword works well to nail the balance between realism and stylization</li><li>I trained this to understand Joel Miller and Oberyn Martell, but it didn't pick up The Mandalorian (see prompts on images for getting them)</li><li>SD wants him to have green/hazel eyes, counteract it with "brown eyes" in prompt or "green eyes" in negative prompt</li><li>My example images were all upscaled using Tiled Diffusion</li></ul>

Files with SHA256: 3206472587ffe8c8ba6da2621560401915b200a2853fb1808266ea7912d6ead5