HOME | DD

Description
Ok, this isn't the beginning of the comparison testing between versions of my LoRA, but I couldn't keep myself from posting this. I didn't train the LoRA on models of women sitting at a café, drinking coffee - whether conjoined or not. But when I asked the LoRA to produce an image of conjoined twins drinking coffee at a café, it produced an image which, in three iterations, yielded this final result. During those three iterations, it even created the second cup of coffee on the table (the original picture only had one cup of coffee in the hand of the far twin). There are a few minor artifacts (this picture wasn't retouched, it's exactly what Stable Diffusion produced), but some can't really be removed without photoshopping them out. Either way, I look forward to what else the LoRA can do!

Description
Ok, this isn't the beginning of the comparison testing between versions of my LoRA, but I couldn't keep myself from posting this. I didn't train the LoRA on models of women sitting at a café, drinking coffee - whether conjoined or not. But when I asked the LoRA to produce an image of conjoined twins drinking coffee at a café, it produced an image which, in three iterations, yielded this final result. During those three iterations, it even created the second cup of coffee on the table (the original picture only had one cup of coffee in the hand of the far twin). There are a few minor artifacts (this picture wasn't retouched, it's exactly what Stable Diffusion produced), but some can't really be removed without photoshopping them out. Either way, I look forward to what else the LoRA can do!