Original artwork and all credits go to <a href="https://www.instagram.com/samdoesarts/" rel="ugc" target="_blank">Sam Yang</a> (samdoesarts) 🎨If you liked the models, join my 🎥<a href="http://instagram.com/ai_rtists/" rel="ugc" target="_blank"> insta</a>, or let's talk on discord: JH#7326Training info : <ul><li>Models finetuned using kane's repo : <a href="https://github.com/kanewallmann/Dreambooth-Stable-Diffusion" rel="ugc" target="_blank">https://github.com/kanewallmann/Dreambooth-Stable-Diffusion</a></li><li>622 final images captioned using BLIP</li><li>Using 3090 on<a href="https://runpod.io?ref=pj6hanur" rel="ugc" target="_blank"> runpod.io</a></li></ul>

JH's samdoesarts 

For v5, data is split :<ul><li>face closeups - Reg : a subset of ffhq by nvidia</li><li>portraits - Reg : women digital portraits and photos generated by sd1.5</li><li>zoomed out / landscapes - Reg : generated by sd1.5 using BLIP captions from training data</li></ul>

JH's samdoesarts v5

JH's samdoesarts v5