LoRA trained on Gio Scotti. I've found it works best with RealisticVision.All images are untouched generations- no img2img, controlnet, photoshop, or upscaling.For reproducibility, use a VAE and Clip Skip 2.Comparison between Revision1 and Revision 2.Revision1: <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/9ebb1dee-eb96-483a-88a6-a60aaf95a658/width=525/9ebb1dee-eb96-483a-88a6-a60aaf95a658.jpeg" />Revision2: <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/80c6e3ac-8ea9-46d2-93e6-c2dfaef4dff8/width=525/80c6e3ac-8ea9-46d2-93e6-c2dfaef4dff8.jpeg" />

Gio Scotti 

Trained on 23 hand-captioned imaged.

Files with SHA256: 71480aed5d7b4f8b1d28890a8a00b6cd3c13b03206404ad72a46752643f16b5a