A LoRA I made for Raven from DC comics. I used <lora:R3333V3N:1> for these pictures but play around with it and see how it works. I have only tested a bit. It works better if you use Danbooru tags. I use the extension Booru tag autocompletion It is available in the extensions tab. This really helps do the prompting for me. I used AbyssOrangeMix2 it does best in anime based models. The trigger words are the best string I came up with to make the LoRA work. This is my first LoRA so bare with me. It was Trained on Anything V3. The hands don't always want to come out nice but you can inpaint them fairly easily. I used abyssorangemix2NSFW I had the best results with it Let me know how you like this one or if I should do more. Leave a rating if you enjoyed it or if you didn't.
I also used https://civitai.com/models/4845/sxz-bloom as a hypernetwork at .5 for these images. Again I used some of @theally embeddings to give some of the images some flair. Her embeddings are next level.
Also huge shoutout to @elldreth for teaching me some new things and giving me the confidence to keep making content. I was getting disheartened with my embeddings and he taught me so much. SO huge shoutout to him for being awesome!!!
|