If you have an idea for a LoRA please reach out, I could use the practice!
This is not a good LoRA. It's totally cursed at full strength, and it requires extra prompting and negative prompting to get satisfying results. That being said, I'll never actually share a LoRA if I wait for perfect results. I'm hoping at least a few people find it useful.
Images are generated with DatAss Rev 3 - Pony Realism
DPM++2M
Karras
Steps 20
CFG 6
triggers: drAsian, asian, dark skin, brown skin
I did often find undesired results that were fixable with prompting. Like I said, this is not a good LoRA
I use adetailer and controlnet for my results, as well as upscaling. I didn't use anything that I think would do the LoRA's job.