A lora about Raven...
I will release a better version later!
this is not that high quality of a lora in comparison to my Jessie lora XL but still works if you want it to but it will get broken face many times and is not flexible to what I would prefer-
some prompt examples-
close up anime style image of ttrvn character wearing a clown costume
full body image of 3d model ttrvn character in a bikini by the beach
don't use ttrvn character without specifying a style! it will give you broken mess otherwise
Background on why this isn't as efficient as my previous Loras-
the bloated dataset and the captions being the way they are probably didn't match each other really well, the new dataset and captions will be more like my Jessie XL lora and hopefully the Lora would be better than this
I couldnt even get that many previews due to colab webui ban for free users so had to code everything in diffusers myself just to generate and find the most suitable epoch... here is the git repo to the colab I created from scratch which is my last resort of using stable diffusion.... https://github.com/ElanElanElanElan/DiffusingColab.git
coding takes time... so that is why I am being delayed on creating Loras... cant do everything