https://github.com/deepghs/waifuc/tree/main
The model data source is derived from Waifuc, a highly robust tool for crawling, filtering, and processing training sets. With the aid of the Waifuc tool and a single 5600g machine, it is possible to efficiently crawl and process over 3000 training images within a span of 2 hours. Training the character "Lora" on this extensive dataset yields significant performance advantages compared to situations with limited data availability.
Trigger Words in page,other tag in lora metainfo.
recommand use model: Kohaku_XL https://civarchive.com/models/162577/kohaku-xl-beta
weight: 0.8~1
Description
FAQ
Comments (5)
Can you make an updated one for Animagine or PonyDiffusion?
without this plan.I believe new base model can generate all most all characters directly like nai3 use none lora model. Please wait base model update.Training character lora is a senseless behavior now,maybe I will try to training cascade model
@kitarz the model seems to be worse when using Animagine and models using it as the base model. PonyDiffusion does not support any loras that's not directly trained on it iirc. Besides, I believe KohakuXL is starting to become outdated, as other models seems to understand prompts way better (might be wrong, though)
@CruxUnhinged At least during my training of this lora, Kohaku was still the best XL anime model. Even though Animagine3 is so fantasy ,Animagine3 is actually a demo , and I believe Animagine V4 and PONY V7 are WIP. The newest of Kohaku XL is training now.
it's not like novelai1.5 release. XL anime model need more and more ft. . without a doubt, lora can't use like nai1.5 can good use for all model.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.