This Lora model is brutal. I have learned the maximum of 1000 pictures. And every picture is handpicked again. We have 1000 different faces in the Lora, different hair colors and lengths, different nations. All pictures have a name: Portrail, and CloseUp.
The Lora is so good that with a value of 0.35 you no longer need any correction in the eyes and face.
Description
FAQ
Comments (4)
The description alone sounds impressive! I just looked at the captions and it says \"girl\": 998, ... so you might have two traps in there, just saying. No hard feelings. :D
Yeah I dunno if I'm gonna take much at face value (no pun) whenever the examples also have "((((Ultra-HD-details, 60+fps-details))))" in the prompt like that's a thing
@zekeverlasting Of course, it was merely meant as a joke. It's nothing but the auto-captioning failing to correctly tag in those cases.
Yeah, prompts are full of bs quite often :D
Like people doubling the refererred resolution would result in any quality gain: "(16K), (32K), (64K), (OVER 9000K)" that kind of nonsense
I don't understand the description (was that machine translated?..), and "no longer need corrections on the eyes and faces" is a bold, bold claim with these examples.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.