Shivam Dreambooth of British actress Jacqueline Bisset, using 238 ref images and 1400 (real) reg images, for 20,000 steps.

This model was trained on particularly high-quality data, with a wide range of poses, and is likely to perform well in pure T2I workflows, which is unusual, for my models:

To be honest, though this model has rich retro quality which I can never get into my dedicated LoRA trainings (this one was extracted from a DreamBooth model), and though it will give you a half-decent face even if the face is very small in frame (which is quite unusual), it really doesn't obey any but the broadest commands.

That said, it is capable sometimes of giving one-shot photoreal renders that need no further work; not many of my models can pull that off.

In any case, I intend to do a V2 of this LoRA, a dedicated LoRA training that will, sadly, lack the retro richness of this one, but actually do what you want; the ability to label things in LoRA training is the key to this, and it's a tricky or impossible business with DreamBooth.

I generally use these models in complex workflows, inpainting faces after initial T2Is, and using ControlNet extensively. So if you're hoping for one-click prompt magic, my models aren't data-curated with this in mind, but rather as tools for traditional workflows that use Photoshop and other older methods.

This model is very likely to produce NSFW and nude renderings unless counter-prompted.