This is a model based on the influencer Sophia Diamond.
Description
FAQ
Comments (8)
😍 👍🏻
Nice work. can you tell me a little about how you train these? What's your average dataset look like? Quantity + size?
https://civitai.com/articles/9798/training-a-lora-for-hunyuan-video-on-windows
So I use this workflow for making LORA. As for my data sets, I usually use between 20 and 30 images and I try to make sure they show the face as much as possible in different angles and lighting. Showing the body is good too as you get the right proportions but you want to prioritise the face more than anything else.
@miniladyAI ok cool. thanks. I am also using that workflow. currently training on a dataset with 135 images which is probably waaay too much. but mine is more of a concept than a character.
thanks for the reply though.
@QualityControl you're very welcome, I haven't tried to train a concept yet but I would guess the requirements are a little different.
any plans to train this for WAN? 👀
I am looking into making WAN lora but my tests so far have been unsuccessful and I won't be around my PC for a couple of weeks but I plan to take another look once I'm back.
@miniladyAI Gotcha!
Would love to see one for WAN, this LoRA is by far the most accurate one, even among the Flux and SDXL lora's. You're doing a great job! :)
I'll be patiently waiting 🫡♥️