👋 Hi! Saved on Civitai to manage my Loras 📝 easily.
The model was extracted from Pixar Style Model checkpoint KittyKnight.
💡Pro tip: Use triggerword in 1.25!
🚫But beware: Using a VAE may not be recommended for this project. I suggest doing some tests with and without a VAE to see what works best for you.
🌟Don't forget to share your amazing artwork! We're always looking for fresh and inspiring designs to showcase. Let's keep the creative juices flowing! 🎨
Description
FAQ
Comments (8)
Great job ! Can you add a LORA for the DreamWorks productions ?
mmmm I can try, this is an extracted model, the one from dreamworks would be an original model, it would just take a while because I have a bit of a perfectionism problem until I accept that there is no way to be perfect and leave it the way it is 🤣
Ok, do what you can, no pressure. Just a suggestion.
which extract method did you use?
Kohya GUI with Extract Lora and Network Dimension in 128
It works fine in 512*512 res, but once switched to which 1024*1024 res it doesn't work
trained on 512, for 1024 need to use Hires. fix
@baiteryamato use Hires. fix，all the parameters are the same but the picture results are not the same,the difference is very big