jp-MilaBrunetta-1700 is the next evolution in my journey of learning Textual Inversions. I went back to giving them names as just letters and numbers became confusing for me.

She was made with stock images of two models with brown hair and similar body shapes to give the results you see.

The specs...

19 images trained for 1700 steps.

12 vectors per token.

Learning rate of .004

Batch size of 4 with 1 gradient accumulation step.