I dont often generate women in SD1.5 so you can most likely get better results than me but i used realistic vision in the sample images. This is trained on the base sd1.5.