I think the name speaks for itself.
Description
The base model is Suzumehachi.
FAQ
Comments (6)
This is better than mines XD i think
nice, well done
Super awesome model! 🥳 How many training pictures did you use to train that?
I don't really remember, but I assume that was around 50 pics per LoRA, 4 LoRAs on different datasets, and after that, I mixed them with various factors until I was satisfied
zayaJun 4, 2023
@dobrosketchkun Wow that's a little less than I expected. Thanks for your reply!🥰 I also want to know if using only 1 of the 4 Loras you trained, would it work well?
@zaya You can try, but I tend to overfit LoRAs a little, so the first-order ones are kind of burned a little, but when they are combined those overfit and burnout dissipate.
@dobrosketchkun Okay, got it! Thanks!!