As a beginner, I made this because I didn't find lora about it.
I'm trying to improve it.(But it has failed until now XD)
It is still not stable,especially in pov.Just generate more.
Setting weight 1.0 is ok.
In the example figure, I used some characters Loras. If there is any offense, please contact me to delete them
Description
FAQ
Comments (11)
Checkpoint? Also images with generation data would be nice
I added the prompts,thanks for you advice.
The model is merged by myself
mainly abyss orange hardcore+pastel
Doesnt work with character LORAs
You can try to adjust the weight, reduce the character's, and increase this. In fact, some characters Lora are really difficult to adapt to this Lora. I have tried several. Some of them perform well, and some of them do not. I added an example of using Ganyu lora
@Saya i tried over 10 character LORAs i had. Didnt work with any of them.
@Wizi Sorry, I don't know. It may really be a coincidence. I have tested 7 character LORAs, two of which can hardly be used with this, and the others have been successfully implemented. I have put them in the example diagram for you to refer to.
@Saya when i say work i mean actually work without changing the artstyle too much. Your example pictures have warped eyes and lose a lot of the details for instance. Which should be avoidable.
thanks for your share. very low chance to work..
I am already considering remaking these old LORAs
@Saya would you say you've learnt a lot since first starting to create these LoRAs?
@weroxig Maybe just some experiential improvements. Making lora is not a technical job, but depends on whether you carefully collect, select and edit the training set. As for the parameters, I think most people who make lora will have their own experience soon