2nd Bounty competition active! The best image gen will win - click here!

Fashion Model Series: #1

My first ever attempt at a LoRA, and decided to make one for the gorgeous Bridget Gao-Hollitt (@reversegaogurl).

I'm still learning so I'd highly appreciate any feedback, so if you like this LoRA, please let me know by leaving a review or clicking the Like buttons. I would love to see the images you create, so please share on here or reach me on Discord (zombiehead91)!

Warning: Due to the training data (modeling/bikini/fashion pics), might produce nudity if not appropriately prompted. Please be respectful of the subject and Civitai terms of use, and avoid any NSFW content. Happy to discuss this further on Discord if unclear...

About the subject:

Bridgett Gao-Hollitt is a Chinese-Australian actress, model, and singer-songwriter who is best known for her work on Super Pumped: The Battle For Uber (2022), and the Australian tv series Home and Away (1988). Her modelling resume includes collaborations with well-known fashion brands such as AKA, AFends, Bassike, Palm Swimwear, Maggie Marilyn, and Worn For Good. She's also an influencer having worked with brands like Tiffany and Co, YSL, Marc Jacobs, L’Oréal, Lacoste, Fluid Form Pilates, Aveda, and Viktoria & Woods.

She's also a singer-songwriter, having released an EP called how was that for you, where (and I swear I'm not making this up) she has a track called ‘Algorithm Defined Fantasy Girl’

Trigger word : ZH-BGHollitt

Other useful triggers: short hair, black hair, long hair, brown hair (see tags in LoRA for full list). Monochrome and greyscale were also intentionally included as triggers (just add em to the negative prompt if you'd rather not have them)

About the sample images

Sample images generated using hi-res.fix WITHOUT ADetailer as a demonstration of the LoRA with minimal prompting. Cherry-picked as a workaround for Civitai TOS, the images you'll get will be a lot closer to the training data.ADetailer use is recommended.
Suggested settings : 0.9 to 1, Clip Skip: 1
Method: DPM++SDE Karras, Steps:30, CFG : 6-7, 512x768 + high res fix with 2x upscale (R-ESRGAN 4x+)

About the training data:

This LoRA was trained on 110 selected images from publicly-available photos (modeling shoots primarily).

Please do not share this LoRA anywhere else without my knowledge, or for any commercial use.