attempt to create ukiyo-e lora for pony
the v1 o give the data set of 768 images if someone want try create a better version
the v2 i use 1000 new images and improve the triggers, works very good with another loras experiment by yourself
the experimental same dataset used in v2 but trained with kohya_ss google colab, the quality is low
the v3 version i will try to release end of june init of july
triggers: ukiyo-e, UKIYO-E
Description
this is a experimental versionif you want better result use the version v2
FAQ
Comments (2)
Really interesting Lora, thank you for sharing, it has a very cool impact on realistic images that I was not expecting
thanks soo much for the comment i love your pic too