Rachel is a new iteration, trying to create a model with less prompting necessary for the model to look right.

Before you get confused, there are two model files for this character LoRa.

She's trained on a dataset of about 70 images at 4000 steps for the standard Rachel Model, and for that you should just use the trigger word Raaachel.

The Rachel K model was done with a different set of triggers, this is to test how these models with the same dataset work with different triggers. In my opinion, this one better captures the full voluptuous body she should have.

For Rachel K, use the trigger words: Raaachel Photograph of an Asian woman with fair skin and black hair, styled in a loose updo.

Or use: Raaachel Photograph of an Asian woman with fair skin and black hair.

If you like her, please post pictures, leave a review, or visit my Ko-Fi.