This lora will give style similar to Dalle-3 for female related pictures.
Dataset: https://huggingface.co/datasets/ProGamerGov/synthetic-dataset-1m-dalle3-high-quality-captions
For data-000051-dalle-female style:
I have chosen female/woman and some variants like goddess manually and with a script, while editing few of the captions.
(Dataset details shows that some images might be from midjourney and good stable diffusion outputs.)
Notes: I didn't test extensively. The images here are mostly with SDXL 1.0 vae 9.0, with some exceptions. You can check the comments and if there's no mention of the model name, it's likely normal SDXL. This is the 3rd epoch of the total 10 epoch output.
The caption prefix: data-000051-dalle-female style
(Use it according to the image as it could produce better or worst results. Try accordingly.)
Tested Lora strength: 0.1-1.0
Clip strength: Not tested.
Clip skip: Tested with 2