Verified: a year ago
SafeTensor
Type | |
Stats | 276 |
Reviews | (58) |
Published | Jul 3, 2023 |
Base Model | |
Trigger Words | Kyungsoo |
Hash | AutoV2 BA0A2C96D1 |
this is a beta version of lora and it really needs a lot of tweaking and this is how i control it when generating
Prompt:
-the required prompt is Kyungsoo but having or not this prompt doesn't affect the image too much.
-face adjustment:If you don't use any prompt for emote description, it will default to cool face. i usually use prompt smile,happy.
-about the outfit: if you just randomize the photo without using the outfit prompt, it will default to a jean jacket and jeans and if you want to change his clothes please describe the clothes carefully and use :( ) to pool the clothing descriptions you want.
-generate:batch count of 3 or more photos when generating photos because you can choose 1 out of 3 photos out of 3 times.the number of steps should be less than 30 is ok