Sophie Rain is an instagram and onlyfans model.

This is a 120 step TI from a 15 image dataset using JernauGurgeh's method of training embeddings. This was trained primarily using close-up or chest-up images, with 5 of them showing most of her body.

How to create a good prompt using this TI

I recommend Jernau's checkpoint and starting with this phrase specifically

"photo of s0phierain-120, a woman"

Extra descriptors include nose ring, necklaces, and open mouth.

If you have feedback about any limitations or suggestions to improve this, let me know. Personally, I find dreambooth fine-tuned models merged with your favorite checkpoint to be more effective than LoRAs or embeddings, but I figured I would share this since I finally found success with training a TI.