I am currently experimenting with the LoRa learning settings for Flux, so here are the experimental models that will be updated in the future.
Description
Personal Ami style
FAQ
Comments (7)
Which means ‘GLSHS’?
I love this LoRA for Flux, it's one of the only ones I've found so far that actually works. My only complaint is that it seems to have an effect where the prompts won't listen to facial expression with this LoRA at a high weight. For example, I was trying to get fangs and an open mouth and it refuses to do it, I removed the LoRA and it did it no problem. But other than that, it's gorgeous and I will get much use out of it!!
Absolutely phenomenal work
So what's the deal here? Did you dramatically overtrain this Lora to the point where it is just spitting out images basically identical to the training data? Or did you post David Dubnitsky's photos and pretend that your Lora created them?
Because if you go to his website, and scroll down to the 11th photo (this one https://1x.com/photo/1213728), that's a different take of one of the photos you have posted below as an example of your Lora's generations.
https://1x.com/dubnitskiy
Explain what's going on.
Overall I love the style of this LoRa. That being said it does have some significant limitations. It seems to only understand skinny or obese, there is no middle ground. Also it's got exactly 1 type of nipple/areola. It also generally ignores age, I tried a variety of prompts between 18-65. But overall this is really a nice LoRa and I appreciate your work.
Seems to be nice.But I have noclue what the different styles do.Why is it so uncommon here to give some information?
Will you add same styles like in pony ?


