<p>Embedding trained on 30 images of actress Ana de Armas. I know, there are already several TIs, LORAs, hypernetworks, what have you, of this actress already (including one of mine). This is just to replace my older hypernetwork if anyone is interested. It was trained on mostly the same dataset, but I did remove a couple of pictures that had an outfit that negatively affected the output from the HN. If anyone is curious, it wa the banded metal dress from the UK premiere of The Gray Man. (Not the LA premiere, that dress was fine). In the older HN, it would show up as vertical stripes on all kinds of outfits as well as a shimmering rainbow/oil slick coloring effect on clothes, even when it didn't have the metal banding.</p>

Files with SHA256: d75e03b9321bcd58245909e18aa7479afb71a3a9b1ebc9bec194453f5c67c1f5